Development and Validation of a Prediction Model for Coronary Artery Disease in Chest Pain Patients: A Real-World Multicenter Study Based on Machine Learning

Development and Validation of a Prediction Model for Coronary Artery Disease in Chest Pain Patients: A Real-World Multicenter Study Based on Machine Learning | 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 Article Development and Validation of a Prediction Model for Coronary Artery Disease in Chest Pain Patients: A Real-World Multicenter Study Based on Machine Learning Songtao He, Jie Jian, Gaiqin Liu, Yuxuan Huang, Longcong Chen This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5234204/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Coronary artery disease (CAD) is a prevalent condition among chest pain patients, and accurate prediction of the disease is crucial to ensure timely interventions and improve patient outcomes. We aim to elaborate a prediction model for CAD in chest pain patients using machine learning approaches. A retrospective analysis was performed using electronic health records of patients who presented with chest pain at seven hospitals. A total of 8474 patients were included in the study, where 63.25% were diagnosed with CAD. The data included demographic information, medical history, and laboratory results. Machine learning algorithms, including Random Forest, CatBoost, XGBoosting, Gradient Boosting, Light Gradient, AdaBoost, Ridge Classifier, Linear Discriminant, Logistic Regression, Decision Tree, SVM, Quadratic Discriminant, K Neighbors, Naive Bayes, and Dummy Classifier were trained and evaluated to predict the presence of CAD.The prediction model achieved an overall accuracy of 0.766 in identifying CAD in chest pain patients. The sensitivity and precision were 0.938 and 0.746, respectively. Important predictors for CAD included age, pulse rate, monocyte, and red cell distribution width SD. The eXtreme Gradient Boosting showed the best performance (area under the receiver operating characteristics, AUROC, 0.820, and 95% CI, 0.801–0.839) Additionally, the model demonstrated robust performance in the validation group. This study successfully developed and validated a prediction model for CAD in chest pain patients using machine learning techniques. The model exhibited good predictive ability and could aid in the early identification of CAD in clinical practice, potentially leading to appropriate interventions and improved patient outcomes. Health sciences/Diseases/Cardiovascular diseases Health sciences/Risk factors Coronary artery disease Chest pain prediction model machine learning Multicenter study Figures Figure 1 Figure 2 Figure 3 1. Introduction The coronary arteries transport blood to the heart muscle and supply it with the necessary ingredients for its function. The term 'coronary artery disease (CAD)' describes the narrowing of these arteries caused by the accumulation of atherosclerotic material in their lumen. This stenosis results in inadequate blood supply to the heart muscle, especially in situations where it has increased needs, leading to myocardial ischemia [ 1 ] . Atherosclerotic material is a soft, fatty substance that forms on the inner surface of arteries through interactions with blood elements (cells and coagulation factors) and fats carried by the blood. Over time, atherosclerotic plaque calcifies and hardens [ 2 , 3 ] . Given its high incidence, hospitalization rates, disability, and mortality rates, the early detection of CAD holds paramount significance [ 4 , 5 ] . In the assessment of patients with chest pain, the determination of coronary artery disease (CAD) probability is a critical aspect of clinical practice. Traditional diagnostic modalities for CAD encompass clinical symptoms and signs, electrocardiography, blood tests, echocardiography, nuclear imaging, and coronary angiography. Each method offers distinct advantages and drawbacks in terms of accuracy and suitability for different scenarios [ 6 – 8 ] . To this end, ESC guidelines advocate for evaluating CAD probability based on age, gender, and symptoms, utilizing the updated Diamond–Forrester (D-F) prediction model [ 9 , 10 ] . Subsequently, pre-test probabilities are utilized to determine whether invasive angiography, non-invasive testing, or no further assessment is warranted [ 11 ] . ACC/AHA guidelines, on the other hand, suggest integrating risk factors alongside a combination of the D-F and CASS registry prediction models [ 12 ] . Meanwhile, recent NICE guidelines propose routine evaluation using coronary computed tomography angiography (CCTA) solely for patients exhibiting atypical or typical angina symptoms [ 13 ] . However, recent studies indicate limitations in the predictive performance of traditional CAD models, particularly concerning obstructive CAD [ 14 – 16 ] . Moreover, these models often fail to comprehensively reflect the regulation status of risk factors such as hypertension, diabetes, and lipid abnormalities [ 14 – 16 ] . In the realm of coronary artery disease (CAD) risk prediction models, there has been a proliferation of research efforts. For instance, [Xu ZJ et al., 2018] [ 17 ] employed logistic regression to establish a prediction score and protocol for the preoperative prediction of significant CAD in patients with rheumatic valvular heart disease. Another notable study conducted by [Andrikou, Ioannis, et al., 2018] [ 18 ] utilized the multivariate Cox regression model to determine the role of serum uric acid (SUA) in cardiovascular risk prediction. Additionally, [Cheng, Yuan, et al., 2023] [ 19 ] verify the association between various inflammatory indicators and ICU mortality by multivariate Logistic regression analysis and Cox proportional hazards model. Another subset of research harnesses machine learning models, which leverage large-scale data processing and complex algorithms to identify potential nonlinear relationships and novel predictive factors, such as [Chen Wang et al., 2021] [ 20 ] construct and validate a machine learning model to predict the risk of CAD based on conventional risk factors and lab test data and [Ali Garavand et al., 2022] [ 21 ] compared the performances of different ML algorithms to assess their effectiveness in developing a model for early CAD diagnosis based on clinical examination features. This descriptive study involved analyzing 303 records. However, these studies were limited by small sample sizes, lack of multicenter data, and absence of external validation. Additionally, there is a dearth of dedicated prediction models for chest pain patients in the literature, highlighting an existing gap in research. To overcome these limitations, we propose a new approach that integrates ML algorithms with the Shapley Additive exPlanations (SHAP) framework. This integration aims to create an interpretable and efficient risk prediction model that is specifically designed to assess factors contributing to CAD diagnosis in patients presenting with chest pain. ML is in general superior when handling many variables—especially if there are complex interactions between these variables. While better suited for handling complex datasets, ML approaches often sacrifice interpretability relative to standard statistics [ 22 ] , posing challenges for clinical decision-making. The SHAP framework is a valuable tool for explaining the variables that positively or negatively impact predictive outcomes. [ 23 ] . By combining ML algorithms with the SHAP framework, our model aims to provide clinicians with a better understanding of the reasons behind predictions. This increased interpretability can assist in evaluating disease severity, enabling more informed clinical decisions, and maximizing opportunities for early intervention in patients with chest pain. This innovative approach aims to bridge the gap between advanced predictive modeling and the practical needs of healthcare professionals. It enhances the clinical utility of CAD risk assessment in chest pain patients. 2. Methods 2.1. Research population and data sources Data were obtained from the Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University, and patients with chest pain were screened from the database. The study includes patients admitted for chest pain from January 1, 2012, to June 30, 2023. Ethical approval was obtained from The Ethics Committee of ChongQing Medical University(Ethics Number: 2023044). As a retrospective study, informed consent was waived by The Ethics Committee of ChongQing Medical University. The research design follows the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guidelines. 2.2. Inclusion and exclusion criteria Inclusion Criteria: (1) Cases admitted from January 1, 2012, to May 31, 2023. (2) Presence of chest pain symptoms. Exclusion Criteria: (1) Previously diagnosed with coronary artery disease. (2) History of coronary stent implantation. 2.3. Outcome The central outcome of this study revolves around the definitive diagnosis of coronary artery disease (CAD), including Acute Coronary Syndromes, unstable or stable angina pectoris, and acute myocardial infarction, as defined by the International Classification of Diseases 10th revision(ICD-10) codes I20-I25. 2.4. Feature selection and data preprocessing Based on relevant research and clinical accessibility, we compiled a dataset encompassing 45 clinical features and predictive factors associated with CAD. These variables include demographic information such as age and gender, comorbidities like hypertension and diabetes, as well as laboratory test indicators including Aspartate aminotransferase (AST), Low-Density Lipoprotein Cholesterol (LDL-C), and High-Density Lipoprotein Cholesterol (HDL-C). All data for the included variables were extracted from the electronic health records of hospitalized patients. The proportions of missing data for all features were less than 30% (Table S1). For features with missing values, we employed the weighted k-nearest neighbors imputation (KNNimpute) method. The fundamental principle of KNNimpute involves identifying K samples in the dataset with Euclidean distances close to the missing values and subsequently utilizing the weighted average of these K samples as the estimated value for the missing data [24] . Feature selection was conducted using the recursive feature elimination algorithm (RFE) based on CatBoost, coupled with 5-fold cross-validation (CV). The core idea of RFE is to construct a model, select the best features, and then iterate through the remaining features until all features have been traversed [25] . 2.5. Model development, evaluation, validation, and interpretation This study utilized data from six medical centers, with one center designated for external validation and the remaining five centers allocated for model development and validation. A stratified random sampling approach was employed to randomly partition the dataset into a training set (70% of subjects) and a validation set (30% of subjects). The former was utilized for the development of machine learning models, while the latter was employed to assess the predictive performance of the models. The study systematically compared the performance of 15 machine learning algorithms with hyperparameter optimization to identify candidate algorithms for predicting CAD in chest pain patients. The selection process was based on the area under the ROC curve (AUC) and accuracy metrics, with logistic regression serving as the baseline model for comparison. The candidate algorithm selection process utilized the sklearn (version 1.2.2) and PyCaret (version 3.2.0) packages. Scikit-learn is a concise yet powerful Python library for machine learning [26] , while PyCaret is a low-code machine learning package designed to simplify model deployment within a user-friendly Python environment [27, 28] . To address the imbalance between positive and negative samples, a preprocessing step involving random undersampling combined with Synthetic Minority Over-sampling Technique (SMOTE) [29] was applied to the training set. This technique aimed to rectify the imbalance issue in the dataset by addressing both overrepresentation and underrepresentation of positive and negative samples, respectively. In the training set performance metrics including AUC, accuracy, sensitivity, precision, and F1 score were computed on the validation set to assess and compare model performance, and Bootstrap resampling was applied to calculate a 95% confidence interval for the AUC. The selection of the optimal predictive model was based on these metrics, Furthermore, the performance of the five predictive models were evaluated in the external validation set by calculating AUC, accuracy, sensitivity,, precision, and F1 score. Lastly, the SHAP (version 0.43.0) framework was utilized to explain the predictions of the best model. SHAP, proposed by Lundberg and Lee, provides a unified framework for interpreting machine learning predictions by assigning an importance value to each feature for a specific prediction. It represents a novel approach to interpreting various black-box machine learning models, and its interpretability performance has been previously validated [30] . 2.6. Statistical analysis All analyses and computations were conducted using Python v3.9.13 and R v4.1.2. This study conducted a comparative analysis of baseline characteristics among chest pain patients in the CAD and non-CAD groups. Categorical variables were presented as frequency (percentage), and intergroup comparisons were executed using the chi-square test. Continuous variables were reported as median (median ± interquartile range). For continuous variables with a normal distribution, the t-test was applied, while the Mann-Whitney U test was employed for those with a skewed distribution. Statistical significance was established at P < 0.05 3. Results 3.1. Patient Characteristics A total of 8474 patients were enrolled in the study, with 6562 of these cases being used for machine learning training and internal validation. The specific selection process is shown in Fig. S1. The main characteristics of patients in the training and internal validation group, are presented in Table 1. Median age of patients with CAD in the training and internal validation groups is 67 (59 -75) years, higher than that of patients without CAD, who are 58 (45-69) years. 3,100 (40.0%) internal-validated patients were female, whereas 613 (37.1%) external-validated patients were female. Table 2 presents a comparison of the main characteristics of patients between the model set and the external validation groups. The median age of patients in the model set is 64 (53-74) years, whereas in the external validation group, it is 65 (54-73) years. The proportion of female patients is 40.0% (2628 patients) in the model set, while in the external validation group, it is 36.0% (689 patients). TABLE 1 | Main characteristics of patients in the training and internal validation groups. Characteristic Overall Non-CAD CAD p value N=6562 N=2535 N=4027 Gender (%) Male 3934 (60.0) 1494 (58.9) 2440 (60.6) 0.191 Female 2628 (40.0) 1041 (41.1) 1587 (39.4) Age 64 [53, 74] 58 [45, 69] 67 [59, 75] <0.001 Transfusion(%) No 6297 (96.0) 2440 (96.3) 3857 (95.8) 0.376 Yes 265 ( 4.0) 95 ( 3.7) 170 ( 4.2) Surgery (%) No 4311 (69.6) 1695 (71.0) 2616 (68.7) 0.067 Yes 1885 (30.4) 694 (29.0) 1191 (31.3) Smoking (%) No 4135 (64.0) 1631 (66.0) 2504 (62.8) 0.009 Yes 2326 (36.0) 840 (34.0) 1486 (37.2) Alcohol (%) No 4755 (76.2) 1839 (75.6) 2916 (76.5) 0.473 Yes 1489 (23.8) 592 (24.4) 897 (23.5) Weight_Change (%) Lose 169 ( 2.8) 110 ( 4.9) 59 ( 1.5) <0.001 No Change 5921 (97.0) 2126 (94.8) 3795 (98.3) Increase 11 ( 0.2) 6 ( 0.3) 5 ( 0.1) SBP 132 [119, 147] 128 [116, 141] 135[120, 150] <0.001 DBP 79 [70, 88] 78 [70, 86] 80 [70, 88] 0.001 PR 78 [69, 87] 79 [71, 89] 76 [68, 86] <0.001 Temperature 36.50 [36.30, 36.60] 36.50 [36.30, 36.70] 36.50 [36.30, 36.60] <0.001 RR 20 [19, 20] 20 [19, 20] 20 [19, 20] 0.094 Hypertension (%) No 3028 (62.9) 1344 (73.7) 1684 (56.4) <0.001 Yes 1783 (37.1) 480 (26.3) 1303 (43.6) Diabetes (%) No 4827 (89.8) 1747 (91.9) 3080 (88.6) <0.001 Yes 551 (10.2) 153 ( 8.1) 398 (11.4) GGT 24.80 [17.00, 43.48] 23.00 [15.30, 40.58] 26.00 [18.00, 46.00] <0.001 Neut 4.64 [3.41, 6.66] 4.72 [3.39, 7.04] 4.60 [3.41, 6.42] 0.009 LDL-C 2.50 [1.94, 3.15] 2.42 [1.89, 3.03] 2.55 [1.97, 3.20] <0.001 HDL-C 1.19 [0.99, 1.45] 1.23 [1.01, 1.50] 1.18 [0.98, 1.43] <0.001 Mono 0.35 [0.26, 0.49] 0.33 [0.23, 0.46] 0.37 [0.27, 0.50] <0.001 Basophil 0.02 [0.01, 0.03] 0.01 [0.00, 0.02] 0.02 [0.01, 0.03] <0.001 Eos 0.09 [0.04, 0.17] 0.08 [0.03, 0.16] 0.10 [0.04, 0.18] <0.001 AST 23.50 [18.59, 34.80] 22.30 [17.70, 30.90] 24.00 [19.00, 38.00] <0.001 Urea 5.60 [4.54, 6.94] 5.41 [4.34, 6.71] 5.74 [4.68, 7.10] <0.001 Uric Acid 327.75 [267.60, 398.08] 311.55 [254.98, 385.85] 336.00 [277.00, 404.62] <0.001 MCV 91.70 [88.30, 95.10] 90.60 [87.10, 93.97] 92.50 [89.00, 95.70] <0.001 MCHC 332 [324, 340] 333 [325, 340] 331 [323, 340] <0.001 MCH 30.50 [29.40, 31.60] 30.30 [29.10, 31.30] 30.70 [29.60, 31.80] <0.001 TBIL 11.00 [8.20, 14.90] 10.80 [8.10, 14.90] 11.10 [8.20, 15.00] 0.133 TP 68.90 [64.10, 73.20] 69.00 [64.60, 73.60] 68.70 [63.97, 73.00] 0.009 TG 1.31 [0.94, 1.92] 1.20 [0.86, 1.75] 1.34 [0.98, 2.01] <0.001 WBC 6.83 [5.43, 8.86] 6.84 [5.39, 9.16] 6.83 [5.45, 8.73] 0.348 ALB 41.50 [38.40, 44.40] 41.70 [38.40, 44.40] 41.40 [38.40, 44.40] 0.735 DBIL 3.60 [2.50, 5.06] 3.70 [2.60, 5.20] 3.50 [2.40, 5.00] <0.001 ALP 74.90 [61.00, 92.00] 73.30 [60.00, 93.00] 75.70 [61.90, 91.60] 0.034 RDW-SD 44.00 [41.60, 47.10] 43.00 [40.70, 46.00] 44.70 [42.20, 47.60] <0.001 RDW-CV 13.20 [12.70, 13.90] 13.20 [12.60, 13.80] 13.20 [12.80, 13.90] <0.001 Creatinine 69.25 [57.90, 83.70] 66.10 [55.70, 78.60] 71.10 [59.70, 86.30] <0.001 Hb 133 [122, 145] 132 [120, 145] 134 [123, 146] <0.001 PLT 190 [153, 231] 193 [155, 239] 188 [151.50, 226] <0.001 TC 4.43 [3.69, 5.23] 4.36 [3.62, 5.11] 4.45 [3.72, 5.30] 0.001 ALT 21.00 [14.20, 32.80] 19.20 [13.00, 30.40] 22.00 [15.40, 34.00] <0.001 K+ 3.98 [3.70, 4.23] 3.98 [3.73, 4.23] 3.98 [3.70, 4.23] 0.24 Na+ 141.00 [138.80, 143.00] 141.00 [139.10, 143.00] 140.90 [138.60, 143.00] 0.039 Lymph 1.43 [1.04, 1.87] 1.39 [1.01, 1.83] 1.46 [1.07, 1.90] <0.001 Abbreviations: CAD, Coronary Artery Disease; SBP, Systolic Blood Pressure; DBP, Diastolic Blood Pressure; PR, Pulse Rate; RR, Respiratory Rate; GGT, Gamma-Glutamyl Transferase; Neut, Neutrophil Count; LDL-C, Low-Density Lipoprotein Cholesterol; HDL-C, High-Density Lipoprotein Cholesterol; Mono, Monocyte Count; Eos, Eosinophil Count; AST, Aspartate Aminotransferase; MCV, Mean Corpuscular Volume; MCHC, Mean Corpuscular Hemoglobin Concentration; MCH, Mean Corpuscular Hemoglobin; TBIL, Total Bilirubin; TP, Total Protein; TG, Triglycerides; WBC, White Blood Cell Count; ALB, Albumin; DBIL, Direct Bilirubin; ALP, Alkaline Phosphatase; RDW-SD, Red Cell Distribution Width SD; RDW-CV, Red Cell Distribution Width CV; Hb, Hemoglobin; PLT, Platelet Count; TC, Total Cholesterol; ALT, Alanine Aminotransferase; Lymph, Lymphocyte Count. TABLE 2 | Comparison of Patient Characteristics between the Model Set and External Validation Groups. Characteristic Overall Model Set Validate Set p value N=8474 N=6562 N=1912 CAD (%) No 3114 (36.7) 2535 ( 38.6) 579 ( 30.3) <0.001 Yes 5360 (63.3) 4027 ( 61.4) 1333 ( 69.7) Age 65 [53, 73] 64 [53, 74] 65 [54, 73] 0.615 Gender (%) Male 5157 (60.9) 3934 ( 60.0) 1223 ( 64.0) 0.002 Female 3317 (39.1) 2628 ( 40.0) 689 ( 36.0) Transfusion (%) No 8116 (95.8) 6297 ( 96.0) 1819 ( 95.1) 0.13 Yes 358 ( 4.2) 265 ( 4.0) 93 ( 4.9) Surgery (%) No 5495 (69.8) 4311 ( 69.6) 1184 ( 70.4) 0.517 Yes 2382 (30.2) 1885 ( 30.4) 497 ( 29.6) Smoking (%) No 4989 (60.4) 4135 ( 64.0) 854 ( 47.6) <0.001 Yes 3266 (39.6) 2326 ( 36.0) 940 ( 52.4) Alcohol (%) No 5851 (72.9) 4755 ( 76.2) 1096 ( 61.6) <0.001 Yes 2171 (27.1) 1489 ( 23.8) 682 ( 38.4) Weight_Change (%) Lose 186 ( 2.8) 169 ( 2.8) 17 ( 3.4) 0.453 No Change 6402 (97.0) 5921 ( 97.0) 481 ( 96.6) Increase 11 ( 0.2) 11 ( 0.2) 0 ( 0.0) SBP 131 [118, 147] 132 [119, 147] 130 [115, 146] <0.001 DBP 79 [70, 88] 79 [70, 88] 80 [70, 90] 0.008 PR 78 [70, 88] 78 [69, 87] 80 [70, 92] <0.001 Temperature 36.50 [36.30, 36.70] 36.50 [36.30, 36.60] 36.50 [36.50, 36.70] <0.001 RR 20 [19, 20] 20 [19, 20] 20 [20, 21] <0.001 Hypertension (%) No 3923 (61.7) 3028 ( 62.9) 895 ( 57.9) <0.001 Yes 2435 (38.3) 1783 ( 37.1) 652 ( 42.1) Diabetes (%) No 6291 (88.9) 4827 ( 89.8) 1464 ( 86.4) <0.001 Yes 782 (11.1) 551 ( 10.2) 231 ( 13.6) GGT 25.00 [17.00, 45.00] 24.80 [17.00, 43.48] 27.00 [17.00, 50.00] <0.001 Neut 4.78 [3.47, 7.01] 4.64 [3.41, 6.66] 5.60 [3.80, 8.30] <0.001 LDL-C 2.49 [1.93, 3.13] 2.50 [1.94, 3.15] 2.45 [1.92, 3.03] 0.004 HDL-C 1.19 [0.99, 1.45] 1.19 [0.99, 1.45] 1.19 [0.98, 1.42] 0.132 Mono 0.37 [0.27, 0.52] 0.35 [0.26, 0.49] 0.46 [0.32, 0.66] <0.001 Basophil 0.02 [0.01, 0.03] 0.02 [0.01, 0.03] 0.02 [0.01, 0.03] <0.001 Eos 0.09 [0.04, 0.17] 0.09 [0.04, 0.17] 0.08 [0.03, 0.15] <0.001 AST 24.00 [18.82, 39.00] 23.50 [18.59, 34.80] 29.00 [19.00, 88.50] <0.001 Urea 5.66 [4.58, 7.06] 5.60 [4.54, 6.94] 5.90 [4.60, 7.50] <0.001 Uric Acid 328.80 [268.00, 400.00] 327.75 [267.60, 398.08] 332.00 [269.00, 406.00] 0.059 MCV 92.00 [88.60, 95.50] 91.70 [88.30, 95.10] 93.00 [90.00, 97.00] <0.001 MCHC 331 [322, 339] 332 [324, 340] 325 [317, 333] <0.001 MCH 30.50 [29.40, 31.60] 30.50 [29.40, 31.60] 30.30 [29.20, 31.50] 0.002 TBIL 10.60 [7.60, 14.70] 11.00 [8.20, 14.90] 8.90 [5.00, 13.90] <0.001 TP 67.99 [63.30, 72.60] 68.90 [64.10, 73.20] 64.60 [60.50, 69.70] <0.001 TG 1.31 [0.95, 1.92] 1.31 [0.94, 1.92] 1.31 [0.97, 1.90] 0.732 WBC 7.00 [5.50, 9.20] 6.83 [5.43, 8.86] 7.80 [5.80, 10.30] <0.001 ALB 41.00 [37.80, 44.00] 41.50 [38.40, 44.40] 39.00 [36.20, 42.20] <0.001 DBIL 3.50 [2.40, 5.00] 3.60 [2.50, 5.06] 3.10 [2.20, 4.60] <0.001 ALP 75 [61, 92] 74.90 [61, 92] 75 [61, 92] 0.579 RDW-SD 43.30 [40.10, 46.50] 44.00 [41.60, 47.10] 12.70 [12.20, 13.40] <0.001 RDW-CV 13.20 [12.70, 13.90] 13.20 [12.70, 13.90] 12.90 [12.30, 13.90] <0.001 Creatinine 69.40 [58.00, 84.00] 69.25 [57.90, 83.70] 70.00 [58.00, 85.00] 0.029 Hb 133 [121, 145] 133 [122, 145] 133 [120, 145] 0.156 PLT 190.00 [153.00, 233.00] 190.00 [153.00, 231.00] 193.00 [154.25, 243.00] 0.004 TC 4.41 [3.70, 5.21] 4.43 [3.69, 5.23] 4.35 [3.70, 5.13] 0.08 ALT 22.00 [14.64, 35.00] 21.00 [14.20, 32.80] 25.00 [16.00, 42.50] <0.001 K+ 3.95 [3.70, 4.20] 3.98 [3.70, 4.23] 3.90 [3.60, 4.20] <0.001 Na+ 140.90 [138.60, 143.00] 141.00 [138.80, 143.00] 140.00 [138.00, 142.00] <0.001 Lymph 1.40 [1.01, 1.84] 1.43 [1.04, 1.87] 1.31 [0.95, 1.73] <0.001 3.2. Risk factors The highest performance point was 40 as estimated by CatBoost-based RFE with 5-fold cross-validation (CV) (Fig.S2). However, upon further examination, we selected 21 variables as the optimal point. This decision was based on the fact that the model's performance at 21 variables was nearly identical to that at 40 variables, offering a more parsimonious model with minimal loss in accuracy. 3.3. Candidate algorithm screening The results of 15 algorithms constructed using the scikit-learn package are presented in Table S2. Based on the model's accuracy and AUC, the top-performing four algorithms (with accuracy > 0.753 and AUC > 0.805) selected for constructing the CAD prediction model for chest pain patients are Random Forest, CatBoost, XGBoost, and Gradient Boosting Classifier. In this study, the traditional Logistic Regression is chosen as the baseline for model comparison. 3.4. Prediction effects of different models The corresponding ROC curves for the 5 models were shown in Fig.1. The Random Forest model had the largest AUC of 0.829(95% CI: 0.810–0.848). The AUCs of CatBoost, XGBoost, Gradient Boosting, and Logistic were 0.816 (95% CI: 0.794–0.836), 0.820 (95% CI: 0.801–0.839), 0.817 (95% CI: 0.800–0.836), 0.736(95% CI: 0.711–0.758), respectively. The accuracy, sensitivity, precision, and F1 scores of the 5 models were also calculated (Table 3) to comprehensively evaluate the performance of each model. The AUC of the XGBoost model (0.820 was second only to random forest (0.829), and it had the highest sensitivity (0.938) and the F1 score (0.841) compared with the other 4 models. Table 3. Performance evaluation of five prediction models in internal validation. Models AUC (95% CI) Accuracy Sensitivity Precision F1 Random Forest 0.829 (0.810–0.848) 0.771 0.859 0.787 0.821 CatBoost 0.816 (0.794–0.836) 0.765 0.851 0.785 0.816 XGBoost 0.820 (0.801–0.839) 0.766 0.938 0.746 0.831 Gradient Boosting 0.817 (0.800–0.836) 0.770 0.850 0.791 0.819 Logistic Regression 0.735 (0.711–0.758) 0.699 0.722 0.773 0.746 3.5.External Validation To assess the generality and robustness of the CAD prediction model we developed for patients with chest pain, we performed external validation using a dataset of 1912 patients from another center. Our results showed that although the predictive power of the four prediction models declined slightly in external validation, they maintained good discriminative power in the XGBoost model, with an ROC-AUC of 0.705, indicating good accuracy in distinguishing patients with and without CAD (Figure 2). In addition, the model shows good accuracy and calibration, suggesting accurate risk estimation(Table 4). These findings suggest that our machine learning-based prediction model for CAD retains its predictive ability in an external patient population, underscoring its potential utility in real-world clinical settings. Table 4 presents the performance of five predictive models. XGBoost achieved the highest Accuracy (0.763), Sensitivity (0.931) and F1 score (0.845), indicating its superior ability to identify true positive cases and maintain a balance between Precision and Sensitivity. Although CatBoost showed the highest AUC (0.786) and strong Precision (0.809), its overall performance did not surpass XGBoost when considering the critical balance between Sensitivity and Precision. The AUC of XGBoost was 0.763, which, although not the highest, complements its high Sensitivity and F1 score, reinforcing its reliability as a robust predictive model. Table 4 Performance evaluation of five prediction models in external validation. Models AUC Accuracy Sensitivity Precision F1 Random Forest 0.780 0.747 0.825 0.815 0.820 Gradient Boosting 0.771 0.762 0.864 0.808 0.835 XGBoost 0.763 0.763 0.931 0.774 0.845 CatBoost 0.786 0.734 0.810 0.809 0.809 Logistic Regression 0.701 0.523 0.339 0.827 0.539 3.6. Model interpretation The feature importance rankings of the 4 prediction models are shown in Fig. 2, including Random Forest (A),Gradient Boosting(B), XGBoost (C), and CatBoost (D). The importance scores were calculated using the built-in attributes in different ML algorithms. In these 4 models, the risk factors most associated with CAD in chest pain patients were Age;Mono; and HDL-C, AST, and ALT. This study illustrated how some characteristics influenced CAD in chest pain patients by incorporating the SHAP framework. Fig. 3A shows the risk factors evaluated using the mean absolute SHAP value in the XGBoost prediction model. Fig. 3B shows the features that impact the outcome. In each important feature row, different color dots represented the final impact of the feature on the outcome, where the red dots represented high-risk value and the blue represented low-risk value. Older age and higher Mono, RDW-SD, AST, and TG were associated with a higher predicted probability of CAD in chest pain patients. The SHAP values indicate the prediction-related features of individual patients and the contribution of each feature to the prediction of prevalence. This study presented a sample of 2 individual predictions, with the red bars showing that the listed characteristics increase the risk of CAD in chest pain patients and the blue bars showing that the listed characteristics decrease the risk of CAD. Fig. 3C shows a 60-year-old patient, whose Pulse Rate, Low-Density Lipoprotein Cholesterol, and Triglycerides were 81, 2.51 mmol/L, and 1.14 mmol/L, respectively. The SHAP value was -1.67. Fig. 3D is a 89-year-old patient, whose Albumin, Pulse Rate, and High-Density Lipoprotein Cholesterol were 38.2 g/L, 55 , and 0.73 mmol/L, respectively. The SHAP value was 6.42. 4. Discussion In this study, we constructed and validated an interpretable ML-based prediction model to predict CAD in chest pain patients. Among the five ML prediction models, the XGBoost model showed the best performance during internal validation, with an AUC of 0.820 (0.801–0.839), Sensitivity of 0.938, and an F1 score of 0.831, surpassing the performance of the other four models. Subsequently, during external validation, the XGBoost model exhibited commendable performance with an AUC of 0.763, Sensitivity of 0.931, and an F1 score of 0.845, highlighting its robustness when applied to an independent dataset. While there was a general decline in performance metrics during external validation, which is expected due to the challenges of generalizing to new data, XGBoost retained its predictive power and balance between Precision and Sensitivity better than the other models. This consistent performance across both internal and external validations underscores XGBoost's reliability and effectiveness as the best choice for predicting coronary artery disease in chest pain patients. In recent studies, XGBoost has been commonly used to build predictive models and has shown excellent discriminative power in many studies. Ding, LanPing et al. [31] developed A ML-based approach, showed optimum performance and might help predict HTPR on clopidogrel after PCI and guide clinical decision-making using 9 ML algorithms, and the XGBoost showed the best performance, with an AUC of 0.82, a precision of 0.80, a recall of 0.44, an F1 score of 0.57, and an accuracy of 0.87. Meanwhile, the importance of features based on SHAP values in this study demonstrated that the ML approach can explain key features of CAD in chest pain patients, as well as that visualization of SHAP summary plots and force maps of the XGBoost model can allow clinicians to visualize and understand key features. In general, the contributions of this study are as follows: first, we utilized a comprehensive set of 15 machine learning models and subsequently selected the most effective four for external validation. Other advanced ML knowledge was used in this study, such as missing value filling based on KNN, feature selection based on RFECV, and the random undersampling with SMOTE oversampling technique to address sample imbalance. The results showed that these methods can effectively improve the prediction of the risk of CAD in chest pain patients. Second, it is always challenging to correctly interpret the large arrays of clinical data of predictive models constructed based on ML and visually present the predicted results to clinicians. Therefore, this study applied the SHAP framework to the XGBoost model “black-box” tree integration model to achieve optimal prediction and interpretability. It also helps clinicians better understand the decision-making process of prediction models and facilitates the use of prediction results, instead of blindly trusting the results of algorithms. Furthermore, the predictive model both took into account the key risk factors and visually explained to clinicians the characteristics that were likely to contribute to a higher (or lower) risk of CAD in chest pain patients. As shown in Fig. 3 B, the higher the concentration of age, monocyte count, Aspartate Aminotransferase, and triglyceride level in chest pain patients, the greater the risk of CAD. Aspartate Aminotransferase (AST) is an enzyme primarily found in the liver and heart, playing a crucial role in amino acid metabolism. AST levels are independently positively associated with the risk and severity of premature CAD, suggesting that these enzymes could serve as surrogate markers for cardiovascular risk in this specific group of patients [32] . Recent studies have also indicated that peripheral monocyte count above 0.45 k/uL may be considered as a predictor of significant CAD in symptomatic patients with chronic coronary syndrome [33] . A higher triglyceride level index may be independently associated with a higher incidence of ASCVDs, CAD, and stroke in people without ASCVDs at baseline [34] . Although this study has several strengths, it also suffers from several limitations. Firstly, the observed decrease in the model's performance during external validation could potentially be attributed to the utilization of data from multiple hospitals, which introduces inherent variations in data collection protocols across different institutions. These variations may have led to disparities in the distribution and characteristics of data between the internal and external validation datasets, thereby impacting the model's ability to generalize effectively beyond the training dataset. As information technology advances and policies progress, data standardization across institutions is expected to improve, which would enhance the model's generalizability and reproducibility in real-world healthcare applications. Despite these challenges, the insights gained from internal validation underscore the potential clinical utility of the model and highlight the importance of ongoing efforts to refine predictive models for accurate and reliable clinical decision-making. Secondly, the absence of image data, particularly key echocardiographic parameters unavailable from the Chongqing Medical University Medical Data Platform, represents a limitation. Future work could explore incorporating multimodal data for a more comprehensive analysis. Additionally, the reliance on laboratory test indicators and comorbidities as the primary variables in our study, although essential for ensuring the prediction model's performance and identifying risk factors, resulted in 21 features. This abundance of features may pose challenges in the practical application of the model within a clinical setting. Following this, we further recommend integrating the model into hospital information systems rather than requiring healthcare personnel to independently utilize it. By seamlessly integrating the model into existing HIS infrastructure, healthcare providers can access predictive insights directly within their workflow, without the need for additional training or expertise in data analysis. Moreover, despite these limitations, our model holds promise for various practical applications. For instance, it could serve as a valuable tool for coronary artery disease (CAD) screening in primary care settings and aid in auxiliary diagnosis. Leveraging clinical data from patients, the model can efficiently identify high-risk individuals, offer decision support for clinicians in selecting appropriate diagnostic procedures, and guide the formulation of personalized treatment plans. Particularly in resource-constrained medical environments, the application of our model can assist healthcare providers in optimizing the use of limited resources, thereby improving early CAD detection rates and treatment outcomes. Although further research and validation are warranted, we remain optimistic about the potential clinical utility of our model and anticipate that future developments will validate its effectiveness and reliability in real-world clinical settings. 5. Conclusion In this study, we successfully developed an advanced XGBoosting model for predicting the risk of CAD in individuals experiencing chest pain. Our model demonstrated superior performance compared with traditional approaches. Notably, age, monocyte count, Aspartate Aminotransferase, and triglyceride level emerged as significant risk factors associated with CAD in chest pain patients. The identification of these key factors not only enhances our understanding of the contributors to CAD but also provides valuable insights for more targeted interventions and risk management strategies in clinical settings. As we conclude, this research serves as a foundational step, laying the groundwork for future investigations that seek to further enhance the precision and efficacy of predicting early CAD risk in individuals presenting with chest pain. The findings presented here underscore the potential for continuous advancements in predictive modeling, ultimately contributing to improved patient outcomes and informed clinical decision-making. Declarations Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the author(s) used ChatGTP3.5 in order to improve readability and language. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. Ethics approval and consent to participate The present study complied with the principles of the Declaration of Helsinki and was approved by The Ethics Committee of ChongQing Medical University (Ethics Number: 2023044). Given the retrospective nature of the study on an anonymized database, informed consent was waived by The Ethics Committee of ChongQing Medical University. Availability of data and material The datasets for this study can be found in Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University (https://demo.yiducloud.com.cn). Competing interests The authors declare that they have no conflict of interest. Funding This research was Sponsored by Natural Science Foundation of Chongqing (CSTB2022NSCQ-MSX0837) and the Intelligent Medicine Research Project of Chongqing Medical University (YJSZHYX202212). Acknowledgements All authors thank their respective institutions for their support. We are also thankful to the Natural Science Foundation of Chongqing and the Intelligent Medicine Research Project of Chongqing Medical University for funding this project . The datasets for this study can be found in Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University (https://demo.yiducloud.com.cn). References Pagliaro B R, Cannata F, Stefanini G G and Bolognese L (2020) Myocardial ischemia and coronary disease in heart failure HEART FAIL REV 25(1): 53-65 Medina-Leyte D J, Zepeda-Garcia O, Dominguez-Perez M, Gonzalez-Garrido A, Villarreal-Molina T and Jacobo-Albavera L (2021) Endothelial Dysfunction, Inflammation and Coronary Artery Disease: Potential Biomarkers and Promising Therapeutical Approaches INT J MOL SCI 22(8): Shreya D, Zamora D I, Patel G S, Grossmann I, Rodriguez K, Soni M, Joshi P K, Patel S C and Sange I (2021) Coronary Artery Calcium Score - A Reliable Indicator of Coronary Artery Disease? Cureus 13(12): e20149 Ralapanawa U and Sivakanesan R (2021) Epidemiology and the Magnitude of Coronary Artery Disease and Acute Coronary Syndrome: A Narrative Review JOURNAL OF EPIDEMIOLOGY AND GLOBAL HEALTH 11(2): 169-77 Roth G A, Abate D, Abate K H, Abay S M, Abbafati C, Abbasi N, Abbastabar H,... GBD C D C (2018) Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017 LANCET 392(10159): 1736-88 Beltrame J F, Crea F, Kaski J C, Ogawa H, Ong P, Sechtem U, Shimokawa H, Merz C N B and Coronary V D I (2017) International standardization of diagnostic criteria for vasospastic angina EUR HEART J 38(33): 2565-8 Knuuti J, Wijns W, Saraste A, Capodanno D, Barbato E, Funck-Brentano C, Prescott E,... European S C (2020) 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes The Task Force for the diagnosis and management of chronic coronary syndromes of the European Society of Cardiology (ESC) EUR HEART J 41(3): 407-77 Tamis-Holland J E, Jneid H, Reynolds H R, Agewall S, Brilakis E S, Brown T M, Lerman A,... Council Q C O R (2019) Contemporary Diagnosis and Management of Patients With Myocardial Infarction in the Absence of Obstructive Coronary Artery Disease: A Scientific Statement From the American Heart Association CIRCULATION 139(18): E891-908 Genders T, Steyerberg E W, Alkadhi H, Leschka S, Desbiolles L, Nieman K, Galema T W,... CAD C (2011) A clinical prediction rule for the diagnosis of coronary artery disease: validation, updating, and extension EUR HEART J 32(11): 1316-30 DIAMOND G A and FORRESTER J S (1979) ANALYSIS OF PROBABILITY AS AN AID IN THE CLINICAL-DIAGNOSIS OF CORONARY-ARTERY DISEASE NEW ENGL J MED 300(24): 1350-8 Knuuti J, Wijns W, Saraste A, Capodanno D, Barbato E, Funck-Brentano C, Prescott E,... European S C (2020) 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes The Task Force for the diagnosis and management of chronic coronary syndromes of the European Society of Cardiology (ESC) EUR HEART J 41(3): 407-77 Fihn S D, Gardin J M, Abrams J, Berra K, Blankenship J C, Dallas A P, Douglas P S,... Yancy C W (2012) 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS Guideline for the Diagnosis and Management of Patients With Stable Ischemic Heart Disease J AM COLL CARDIOL 60(24): E44-164 Skinner J S, Smeeth L, Kendall J M, Adams P C, Timmis A and Chest P G D G (2010) NICE guidance. Chest pain of recent onset: assessment and diagnosis of recent onset chest pain or discomfort of suspected cardiac origin HEART 96(12): 974-8 Goldstein B A, Navar A M and Carter R E (2017) Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges EUR HEART J 38(23): 1805-14 Ali M M, Gul S, Naqvi M, Hakam L, Inayat A, Saleem S, Polavarpu M and Syed M A (2021) Utility of Coronary Artery Calcium Scores in Predicting Risk of Subclinical Cardiovascular Atherosclerotic Disease: An Analysis of Limitations to its Adoption With Policy Recommendations. Cureus 13(4): e14647 Doolub G, Mamalakis M, Alabed S, Van der Geest R J, Swift A J, Rodrigues J C L, Garg P, Joshi N V and Dastidar A (2023) Artificial Intelligence as a Diagnostic Tool in Non-Invasive Imaging in the Assessment of Coronary Artery Disease. Medical sciences (Basel, Switzerland) 11(1): Xu Z, Pan J, Chen T, Zhou Q, Wang Q, Cao H, Fan F,... Wang D (2018) A prediction score for significant coronary artery disease in Chinese patients ≥50 years old referred for rheumatic valvular heart disease surgery INTERACT CARDIOV TH 26(4): 623-30 Andrikou I, Tsioufis C, Dimitriadis K, Konstantinidis D, Kasiakogias A, Kouremeti M, Andrikou E,... Tousoulis D (2018) Uric acid as an independent predictor of coronary artery disease in essential hypertension: Data from an 8-year-follow-up study CLIN EXP PHARMACOL P 45(8): 866-9 Cheng Y, Chen Y, Mao M, Wang R, Zhu J and He Q (2023) Association of inflammatory indicators with intensive care unit mortality in critically ill patients with coronary heart disease FRONT IMMUNOL 14( Wang C, Zhao Y, Jin B Y, Gan X D, Liang B, Xiang Y, Zhang X K, Lu Z B and Zheng F (2021) Development and Validation of a Predictive Model for Coronary Artery Disease Using Machine Learning FRONTIERS IN CARDIOVASCULAR MEDICINE 8( Garavand A, Salehnasab C, Behmanesh A, Aslani N, Zadeh A H and Ghaderzadeh M (2022) Efficient Model for Coronary Artery Disease Diagnosis: A Comparative Study of Several Machine Learning Algorithms J HEALTHC ENG 2022( Ley C, Martin R K, Pareek A, Groll A, Seil R and Tischer T (2022) Machine learning and conventional statistics: making sense of the differences KNEE SURG SPORT TR A 30(3): 753-7 Ogami C, Tsuji Y, Seki H, Kawano H, To H, Matsumoto Y and Hosono H (2021) An artificial neural network-pharmacokinetic model and its interpretation using Shapley additive explanations CPT-PHARMACOMETRICS & SYSTEMS PHARMACOLOGY 10(7): 760-8 Liao S G, Lin Y, Kang D D, Chandra D, Bon J, Kaminski N, Sciurba F C and Tseng G C (2014) Missing value imputation in high-dimensional phenomic data: imputable or not, and how? BMC BIOINFORMATICS 15( Chen R, Dewi C, Huang S and Caraka R E (2020) Selecting critical features for data classification based on machine learning methods JOURNAL OF BIG DATA 7(1): Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M,... Duchesnay E (2011) Scikit-learn: Machine Learning in Python J MACH LEARN RES 12(2825-30 Ali M (2020) PyCaret: An open source, low-code machine learning library in Python. Muqeet M, Malik H, Panhwar S, Khan I U, Hussain F, Asghar Z, Khatri Z and Mahar R B (2023) Enhanced cellulose nanofiber mechanical stability through ionic crosslinking and interpretation of adsorption data using machine learning INT J BIOL MACROMOL 237( Fernandez A, Garcia S, Herrera F and Chawla N V (2018) SMOTE for Learning from Imbalanced Data: Progress and Challenges, Marking the 15-year Anniversary J ARTIF INTELL RES 61(863-905 Rodriguez-Perez R and Bajorath J (2020) Interpretation of machine learning models using shapley values: application to compound potency and multi-target activity predictions J COMPUT AID MOL DES 34(10): 1013-26 Ding L P, Li P, Yang L R, Pan M M, Zhou M, Zhang C, Yan Y D,... Gu Z C (2024) A novel machine learning model to predict high on-treatment platelet reactivity on clopidogrel in Asian patients after percutaneous coronary intervention INT J CLIN PHARM-NET 46(1): 90-100 Masoudkabir F, Karbalai S, Vasheghani-Farahani A, Aliabadi L L, Boroumand M A, Aiatollahzade-Esfahani F, Pashing M,... Saadat S (2011) The Association of Liver Transaminase Activity With Presence and Severity of Premature Coronary Artery Disease ANGIOLOGY 62(8): 614-9 Urbanowicz T, Olasinska-Wisniewska A, Michalak M, Komosa A, Filipiak K J, Uruski P, Radziemski A, Tykarski A and Jemielity M (2023) Predictive role of monocyte count for significant coronary artery disease identification in patients with stable coronary artery disease CARDIOL J Ding X B, Wang X Z, Wu J, Zhang M L and Cui M Z (2021) Triglyceride-glucose index and the incidence of atherosclerotic cardiovascular diseases: a meta-analysis of cohort studies CARDIOVASC DIABETOL 20(1): Additional Declarations No competing interests reported. Supplementary Files supply.docx Cite Share Download PDF Status: Posted Version 1 posted 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-5234204","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":374847644,"identity":"b4125e32-c9ac-4480-92a8-5d2690e948de","order_by":0,"name":"Songtao He","email":"","orcid":"","institution":"Chongqing Medical University","correspondingAuthor":false,"prefix":"","firstName":"Songtao","middleName":"","lastName":"He","suffix":""},{"id":374847650,"identity":"01028e5c-2400-4767-b13d-3b4d5a862c39","order_by":1,"name":"Jie Jian","email":"","orcid":"","institution":"Chongqing University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Jian","suffix":""},{"id":374847651,"identity":"525518b1-f465-4e43-af65-866483e35bf2","order_by":2,"name":"Gaiqin Liu","email":"","orcid":"","institution":"Chongqing University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Gaiqin","middleName":"","lastName":"Liu","suffix":""},{"id":374847654,"identity":"29860967-4451-447e-9f55-51a8e754fb80","order_by":3,"name":"Yuxuan Huang","email":"","orcid":"","institution":"Hospital of Yubei District of Chongqing City","correspondingAuthor":false,"prefix":"","firstName":"Yuxuan","middleName":"","lastName":"Huang","suffix":""},{"id":374847658,"identity":"83e6858f-ef4f-487c-922d-2ad6a5a962e6","order_by":4,"name":"Longcong Chen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwUlEQVRIiWNgGAWjYDACdjDJLAflMhOhBaKG2Zh0LYkNRGvhb2Z+wHRzh3V6f3t2mgRDhXViA/vZA3i1SBxmM2DOPZOeO+PM220SDGfSExt48hLwajFg5mFgzm07nLtBInebBGPb4cQGCR4DorSkG4C1/CNBSwJESwMRWmB+MQT6ZbNFwrF04zaeHPxa+NubHzDn7rCW52/P3XjjQ421bD/7GfxagID9B2MDiE4AIwY2QurBAK5lFIyCUTAKRgE2AADmGjz4urWpuAAAAABJRU5ErkJggg==","orcid":"","institution":"Chongqing Medical University","correspondingAuthor":true,"prefix":"","firstName":"Longcong","middleName":"","lastName":"Chen","suffix":""}],"badges":[],"createdAt":"2024-10-09 16:53:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5234204/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5234204/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":69911133,"identity":"f7e89f59-1a96-48be-9780-7d95a3894eae","added_by":"auto","created_at":"2024-11-26 13:53:32","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":53020,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves of five machine learning prediction models in internal validation.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5234204/v1/50ad80a3fc2952ec28a8a7a7.png"},{"id":69911127,"identity":"86a5d6a3-7b61-42c5-be49-7f231e952dd2","added_by":"auto","created_at":"2024-11-26 13:53:31","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":410734,"visible":true,"origin":"","legend":"\u003cp\u003eImportance ranking of features in four CAD prediction algorithms. (A) Gradient Boosting, (B) Random Forest, (C) LightGBM, and (D) CatBoost.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5234204/v1/453d57a6f6aa12a22e466ecf.png"},{"id":69911128,"identity":"19025469-f612-4e81-b56f-41a79a174d7d","added_by":"auto","created_at":"2024-11-26 13:53:31","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":682029,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP interpretation of the Gradient Boosting model. (A) Importance ranking of the model prediction features. (B) The effect of each feature on the model output. In each row, each dot represents a patient, and the color of the dot indicates the feature value: red indicates a larger value and blue indicates a lower value. (C and D) Individualized predictions for two patients. The red and blue bars represent risk factors and protective factors, respectively. Longer bars indicate higher functional importance.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5234204/v1/44c5a6a30a996234a9342e2a.png"},{"id":95525871,"identity":"eb7fe692-fea3-4a3f-9184-1072196aec1a","added_by":"auto","created_at":"2025-11-10 10:05:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2425416,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5234204/v1/ffe66e37-6785-41aa-b522-0b13accbd579.pdf"},{"id":69911135,"identity":"cbb03bf4-cfe8-4f5e-bc89-e15df43b1cfb","added_by":"auto","created_at":"2024-11-26 13:53:33","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":82716,"visible":true,"origin":"","legend":"","description":"","filename":"supply.docx","url":"https://assets-eu.researchsquare.com/files/rs-5234204/v1/0b25bde58d2a643950427a4d.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Prediction Model for Coronary Artery Disease in Chest Pain Patients: A Real-World Multicenter Study Based on Machine Learning","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe coronary arteries transport blood to the heart muscle and supply it with the necessary ingredients for its function. The term 'coronary artery disease (CAD)' describes the narrowing of these arteries caused by the accumulation of atherosclerotic material in their lumen. This stenosis results in inadequate blood supply to the heart muscle, especially in situations where it has increased needs, leading to myocardial ischemia\u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. Atherosclerotic material is a soft, fatty substance that forms on the inner surface of arteries through interactions with blood elements (cells and coagulation factors) and fats carried by the blood. Over time, atherosclerotic plaque calcifies and hardens\u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e. Given its high incidence, hospitalization rates, disability, and mortality rates, the early detection of CAD holds paramount significance \u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the assessment of patients with chest pain, the determination of coronary artery disease (CAD) probability is a critical aspect of clinical practice. Traditional diagnostic modalities for CAD encompass clinical symptoms and signs, electrocardiography, blood tests, echocardiography, nuclear imaging, and coronary angiography. Each method offers distinct advantages and drawbacks in terms of accuracy and suitability for different scenarios \u003csup\u003e[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. To this end, ESC guidelines advocate for evaluating CAD probability based on age, gender, and symptoms, utilizing the updated Diamond\u0026ndash;Forrester (D-F) prediction model\u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. Subsequently, pre-test probabilities are utilized to determine whether invasive angiography, non-invasive testing, or no further assessment is warranted\u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. ACC/AHA guidelines, on the other hand, suggest integrating risk factors alongside a combination of the D-F and CASS registry prediction models\u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Meanwhile, recent NICE guidelines propose routine evaluation using coronary computed tomography angiography (CCTA) solely for patients exhibiting atypical or typical angina symptoms\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. However, recent studies indicate limitations in the predictive performance of traditional CAD models, particularly concerning obstructive CAD\u003csup\u003e[\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e. Moreover, these models often fail to comprehensively reflect the regulation status of risk factors such as hypertension, diabetes, and lipid abnormalities\u003csup\u003e[\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the realm of coronary artery disease (CAD) risk prediction models, there has been a proliferation of research efforts. For instance, [Xu ZJ et al., 2018]\u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e employed logistic regression to establish a prediction score and protocol for the preoperative prediction of significant CAD in patients with rheumatic valvular heart disease. Another notable study conducted by [Andrikou, Ioannis, et al., 2018]\u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e utilized the multivariate Cox regression model to determine the role of serum uric acid (SUA) in cardiovascular risk prediction. Additionally, [Cheng, Yuan, et al., 2023]\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e verify the association between various inflammatory indicators and ICU mortality by multivariate Logistic regression analysis and Cox proportional hazards model. Another subset of research harnesses machine learning models, which leverage large-scale data processing and complex algorithms to identify potential nonlinear relationships and novel predictive factors, such as [Chen Wang et al., 2021]\u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e construct and validate a machine learning model to predict the risk of CAD based on conventional risk factors and lab test data and [Ali Garavand et al., 2022]\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e compared the performances of different ML algorithms to assess their effectiveness in developing a model for early CAD diagnosis based on clinical examination features. This descriptive study involved analyzing 303 records. However, these studies were limited by small sample sizes, lack of multicenter data, and absence of external validation. Additionally, there is a dearth of dedicated prediction models for chest pain patients in the literature, highlighting an existing gap in research.\u003c/p\u003e \u003cp\u003eTo overcome these limitations, we propose a new approach that integrates ML algorithms with the Shapley Additive exPlanations (SHAP) framework. This integration aims to create an interpretable and efficient risk prediction model that is specifically designed to assess factors contributing to CAD diagnosis in patients presenting with chest pain. ML is in general superior when handling many variables\u0026mdash;especially if there are complex interactions between these variables. While better suited for handling complex datasets, ML approaches often sacrifice interpretability relative to standard statistics\u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e, posing challenges for clinical decision-making. The SHAP framework is a valuable tool for explaining the variables that positively or negatively impact predictive outcomes. \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e. By combining ML algorithms with the SHAP framework, our model aims to provide clinicians with a better understanding of the reasons behind predictions. This increased interpretability can assist in evaluating disease severity, enabling more informed clinical decisions, and maximizing opportunities for early intervention in patients with chest pain. This innovative approach aims to bridge the gap between advanced predictive modeling and the practical needs of healthcare professionals. It enhances the clinical utility of CAD risk assessment in chest pain patients.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003ch3\u003e2.1.\u0026nbsp;Research population and data sources\u003c/h3\u003e\n\u003cp\u003eData were obtained from the Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University, and patients with chest pain were screened from the database. The study includes patients admitted for chest pain from January 1, 2012, to June 30, 2023. Ethical approval was obtained from The Ethics Committee of ChongQing Medical University(Ethics Number: 2023044). As a retrospective study, informed consent was waived by The Ethics Committee of ChongQing Medical University. The research design follows the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guidelines.\u003c/p\u003e\n\u003ch3\u003e2.2.\u0026nbsp;Inclusion and exclusion criteria\u003c/h3\u003e\n\u003cp\u003eInclusion Criteria:\u003c/p\u003e\n\u003cp\u003e(1) Cases admitted from January 1, 2012, to May 31, 2023.\u003c/p\u003e\n\u003cp\u003e(2) Presence of chest pain symptoms.\u003c/p\u003e\n\u003cp\u003eExclusion Criteria:\u003c/p\u003e\n\u003cp\u003e(1) Previously diagnosed with coronary artery disease.\u003c/p\u003e\n\u003cp\u003e(2) History of coronary stent implantation.\u003c/p\u003e\n\u003ch3\u003e2.3.\u0026nbsp;Outcome\u003c/h3\u003e\n\u003cp\u003eThe central outcome of this study revolves around the definitive diagnosis of coronary artery disease (CAD), including Acute Coronary Syndromes, unstable or stable angina pectoris, and acute myocardial infarction, as defined by the International Classification of Diseases 10th revision(ICD-10) codes I20-I25.\u003c/p\u003e\n\u003ch3\u003e2.4.\u0026nbsp;Feature selection and data preprocessing\u003c/h3\u003e\n\u003cp\u003eBased on relevant research and clinical accessibility, we compiled a dataset encompassing 45 clinical features and predictive factors associated with CAD. These variables include demographic information such as age and gender, comorbidities like hypertension and diabetes, as well as laboratory test indicators including Aspartate aminotransferase (AST), Low-Density Lipoprotein Cholesterol (LDL-C), and High-Density Lipoprotein Cholesterol (HDL-C). All data for the included variables were extracted from the electronic health records of hospitalized patients. The proportions of missing data for all features were less than 30% (Table S1).\u003c/p\u003e\n\u003cp\u003eFor features with missing values, we employed the weighted k-nearest neighbors imputation (KNNimpute) method. The fundamental principle of KNNimpute involves identifying K samples in the dataset with Euclidean distances close to the missing values and subsequently utilizing the weighted average of these K samples as the estimated value for the missing data\u003csup\u003e[24]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eFeature selection was conducted using the recursive feature elimination algorithm\u0026nbsp;(RFE) based on CatBoost, coupled with 5-fold cross-validation (CV). The core idea of RFE is to construct a model, select the best features, and then iterate through the remaining features until all features have been traversed\u003csup\u003e[25]\u003c/sup\u003e.\u003c/p\u003e\n\u003ch3\u003e2.5.\u0026nbsp;Model development, evaluation,\u0026nbsp;validation,\u0026nbsp;and interpretation\u003c/h3\u003e\n\u003cp\u003eThis study utilized data from six medical centers, with one center designated for external validation and the remaining five centers allocated for model development and validation. A stratified random sampling approach was employed to randomly partition the dataset into a training set (70% of subjects) and a validation set (30% of subjects). The former was utilized for the development of machine learning models, while the latter was employed to assess the predictive performance of the models.\u003c/p\u003e\n\u003cp\u003eThe study systematically compared the performance of 15 machine learning algorithms with hyperparameter optimization to identify candidate algorithms for predicting CAD in chest pain patients. The selection process was based on the area under the ROC curve (AUC) and accuracy metrics, with logistic regression serving as the baseline model for comparison. The candidate algorithm selection process utilized the sklearn (version 1.2.2) and PyCaret (version 3.2.0) packages. Scikit-learn is a concise yet powerful Python library for machine learning\u003csup\u003e[26]\u003c/sup\u003e, while PyCaret is a low-code machine learning package designed to simplify model deployment within a user-friendly Python environment\u003csup\u003e[27, 28]\u003c/sup\u003e. To address the imbalance between positive and negative samples, a preprocessing step involving random undersampling combined with Synthetic Minority Over-sampling Technique (SMOTE)\u0026nbsp;\u003csup\u003e[29]\u003c/sup\u003e was applied to the training set. This technique aimed to rectify the imbalance issue in the dataset by addressing both overrepresentation and underrepresentation of positive and negative samples, respectively.\u003c/p\u003e\n\u003cp\u003eIn the training set performance metrics including AUC, accuracy, sensitivity, precision, and F1 score were computed on the validation set to assess and compare model performance, and Bootstrap resampling was applied to calculate a 95% confidence interval for the AUC. The selection of the optimal predictive model was based on these metrics,\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFurthermore, the performance of the five predictive models were evaluated in the external validation set by calculating AUC, accuracy, sensitivity,, precision, and F1 score.\u003c/p\u003e\n\u003cp\u003eLastly, the SHAP (version 0.43.0) framework was utilized to explain the predictions of the best model. SHAP, proposed by Lundberg and Lee, provides a unified framework for interpreting machine learning predictions by assigning an importance value to each feature for a specific prediction. It represents a novel approach to interpreting various black-box machine learning models, and its interpretability performance has been previously validated\u003csup\u003e[30]\u003c/sup\u003e.\u003c/p\u003e\n\u003ch3\u003e2.6.\u0026nbsp;Statistical analysis\u003c/h3\u003e\n\u003cp\u003eAll analyses and computations were conducted using Python v3.9.13 and R v4.1.2. This study conducted a comparative analysis of baseline characteristics among chest pain patients in the CAD and non-CAD groups. Categorical variables were presented as frequency (percentage), and intergroup comparisons were executed using the chi-square test. Continuous variables were reported as median (median ± interquartile range). For continuous variables with a normal distribution, the t-test was applied, while the Mann-Whitney U test was employed for those with a skewed distribution. Statistical significance was established at P \u0026lt; 0.05\u003c/p\u003e"},{"header":"3. Results","content":"\u003ch3\u003e3.1. Patient Characteristics\u003c/h3\u003e\n\u003cp\u003eA total of 8474 patients were enrolled in the study, with 6562 of these cases being used for machine learning training and internal validation. The specific selection process is shown in Fig. S1. The main characteristics of patients in the training and internal validation group, are presented in Table 1. Median age of patients with CAD in the training and internal validation groups is 67 (59 -75) years, higher than that of patients without CAD, who are 58 (45-69) years. 3,100 (40.0%) internal-validated patients were female, whereas 613 (37.1%) external-validated patients were female. Table 2 presents a comparison of the main characteristics of patients between the model set and the external validation groups. The median age of patients in the model set is 64 (53-74) years, whereas in the external validation group, it is 65 (54-73) years. The proportion of female patients is 40.0% (2628 patients) in the model set, while in the external validation group, it is 36.0% (689 patients).\u003c/p\u003e\n\u003cp\u003eTABLE 1 | Main characteristics of patients in the training and internal validation groups.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"606\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 26.6116%;\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23.8017%;\"\u003e\n \u003cp\u003eNon-CAD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23.4711%;\"\u003e\n \u003cp\u003eCAD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.4132%;\"\u003e\n \u003cp\u003ep value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 15.7025%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 26.6116%;\"\u003e\n \u003cp\u003eN=6562\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23.8017%;\"\u003e\n \u003cp\u003eN=2535\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 23.4711%;\"\u003e\n \u003cp\u003eN=4027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eGender (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e3934 (60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1494 (58.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e2440 (60.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.191\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e2628 (40.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1041 (41.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1587 (39.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eAge\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e64 [53, 74]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e58 [45, 69]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e67 [59, 75]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTransfusion(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e6297 (96.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e2440 (96.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e3857 (95.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.376\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e265 ( 4.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e95 ( 3.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e170 ( 4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eSurgery (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4311 (69.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1695 (71.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e2616 (68.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.067\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1885 (30.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e694 (29.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1191 (31.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eSmoking (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4135 (64.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1631 (66.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e2504 (62.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e2326 (36.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e840 (34.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1486 (37.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eAlcohol (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4755 (76.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1839 (75.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e2916 (76.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.473\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1489 (23.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e592 (24.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e897 (23.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eWeight_Change (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eLose\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e169 ( 2.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e110 ( 4.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e59 ( 1.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo Change\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e5921 (97.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e2126 (94.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e3795 (98.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eIncrease\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e11 ( 0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e6 ( 0.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e5 ( 0.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eSBP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e132 [119, 147]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e128 [116, 141]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e135[120, 150]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eDBP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e79 [70, 88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e78 [70, 86]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e80 [70, 88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003ePR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e78 [69, 87]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e79 [71, 89]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e76 [68, 86]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTemperature\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e36.50 [36.30, 36.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e36.50 [36.30, 36.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e36.50 [36.30, 36.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eRR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e20 [19, 20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e20 [19, 20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e20 [19, 20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.094\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eHypertension (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e3028 (62.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1344 (73.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1684 (56.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1783 (37.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e480 (26.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1303 (43.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eDiabetes (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4827 (89.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1747 (91.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e3080 (88.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e551 (10.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e153 ( 8.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e398 (11.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eGGT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e24.80 [17.00, 43.48]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e23.00 [15.30, 40.58]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e26.00 [18.00, 46.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNeut\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4.64 [3.41, 6.66]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e4.72 [3.39, 7.04]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e4.60 [3.41, 6.42]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eLDL-C\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e2.50 [1.94, 3.15]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e2.42 [1.89, 3.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e2.55 [1.97, 3.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eHDL-C\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1.19 [0.99, 1.45]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1.23 [1.01, 1.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1.18 [0.98, 1.43]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eMono\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e0.35 [0.26, 0.49]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e0.33 [0.23, 0.46]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e0.37 [0.27, 0.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eBasophil\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e0.02 [0.01, 0.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e0.01 [0.00, 0.02]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e0.02 [0.01, 0.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eEos\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e0.09 [0.04, 0.17]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e0.08 [0.03, 0.16]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e0.10 [0.04, 0.18]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eAST\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e23.50 [18.59, 34.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e22.30 [17.70, 30.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e24.00 [19.00, 38.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eUrea\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e5.60 [4.54, 6.94]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e5.41 [4.34, 6.71]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e5.74 [4.68, 7.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eUric Acid\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e327.75 [267.60, 398.08]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e311.55 [254.98, 385.85]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e336.00 [277.00, 404.62]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eMCV\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e91.70 [88.30, 95.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e90.60 [87.10, 93.97]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e92.50 [89.00, 95.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eMCHC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e332 [324, 340]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e333 [325, 340]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e331 [323, 340]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eMCH\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e30.50 [29.40, 31.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e30.30 [29.10, 31.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e30.70 [29.60, 31.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTBIL\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e11.00 [8.20, 14.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e10.80 [8.10, 14.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e11.10 [8.20, 15.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.133\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e68.90 [64.10, 73.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e69.00 [64.60, 73.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e68.70 [63.97, 73.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTG\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1.31 [0.94, 1.92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1.20 [0.86, 1.75]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1.34 [0.98, 2.01]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eWBC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e6.83 [5.43, 8.86]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e6.84 [5.39, 9.16]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e6.83 [5.45, 8.73]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.348\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eALB\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e41.50 [38.40, 44.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e41.70 [38.40, 44.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e41.40 [38.40, 44.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.735\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eDBIL\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e3.60 [2.50, 5.06]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e3.70 [2.60, 5.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e3.50 [2.40, 5.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eALP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e74.90 [61.00, 92.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e73.30 [60.00, 93.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e75.70 [61.90, 91.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.034\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eRDW-SD\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e44.00 [41.60, 47.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e43.00 [40.70, 46.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e44.70 [42.20, 47.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eRDW-CV\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e13.20 [12.70, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e13.20 [12.60, 13.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e13.20 [12.80, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eCreatinine\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e69.25 [57.90, 83.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e66.10 [55.70, 78.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e71.10 [59.70, 86.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eHb\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e133 [122, 145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e132 [120, 145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e134 [123, 146]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003ePLT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e190 [153, 231]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e193 [155, 239]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e188 [151.50, 226]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eTC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e4.43 [3.69, 5.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e4.36 [3.62, 5.11]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e4.45 [3.72, 5.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eALT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e21.00 [14.20, 32.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e19.20 [13.00, 30.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e22.00 [15.40, 34.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eK+\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e3.98 [3.70, 4.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e3.98 [3.73, 4.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e3.98 [3.70, 4.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eNa+\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e141.00 [138.80, 143.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e141.00 [139.10, 143.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e140.90 [138.60, 143.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 15.7025%;\"\u003e\n \u003cp\u003eLymph\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 26.6116%;\"\u003e\n \u003cp\u003e1.43 [1.04, 1.87]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.8017%;\"\u003e\n \u003cp\u003e1.39 [1.01, 1.83]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 23.4711%;\"\u003e\n \u003cp\u003e1.46 [1.07, 1.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.4132%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCAD, Coronary Artery Disease; SBP, Systolic Blood Pressure; DBP, Diastolic Blood Pressure; PR, Pulse Rate; RR, Respiratory Rate; GGT, Gamma-Glutamyl Transferase; Neut, Neutrophil Count; LDL-C, Low-Density Lipoprotein Cholesterol; HDL-C, High-Density Lipoprotein Cholesterol; Mono, Monocyte Count; Eos, Eosinophil Count; AST, Aspartate Aminotransferase; MCV, Mean Corpuscular Volume; MCHC, Mean Corpuscular Hemoglobin Concentration; MCH, Mean Corpuscular Hemoglobin; TBIL, Total Bilirubin; TP, Total Protein; TG, Triglycerides; WBC, White Blood Cell Count; ALB, Albumin; DBIL, Direct Bilirubin; ALP, Alkaline Phosphatase; RDW-SD, Red Cell Distribution Width SD; RDW-CV, Red Cell Distribution Width CV; Hb, Hemoglobin; PLT, Platelet Count; TC, Total Cholesterol; ALT, Alanine Aminotransferase; Lymph, Lymphocyte Count.\u003c/p\u003e\n\u003cp\u003eTABLE 2 | Comparison of Patient Characteristics between the Model Set and External Validation Groups.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"606\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003eModel Set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003eValidate Set\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003ep value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003eN=8474\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003eN=6562\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003eN=1912\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eCAD (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3114 (36.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e2535 ( 38.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e579 ( 30.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\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: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e5360 (63.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4027 ( 61.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1333 ( 69.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eAge\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e65 [53, 73]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e64 [53, 74]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e65 [54, 73]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.615\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eGender (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e5157 (60.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e3934 ( 60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1223 ( 64.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3317 (39.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e2628 ( 40.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e689 ( 36.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTransfusion (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e8116 (95.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e6297 ( 96.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1819 ( 95.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e358 ( 4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e265 ( 4.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e93 ( 4.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eSurgery (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e5495 (69.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4311 ( 69.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1184 ( 70.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.517\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e2382 (30.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1885 ( 30.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e497 ( 29.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eSmoking (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e4989 (60.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4135 ( 64.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e854 ( 47.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3266 (39.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e2326 ( 36.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e940 ( 52.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eAlcohol (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e5851 (72.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4755 ( 76.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1096 ( 61.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e2171 (27.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1489 ( 23.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e682 ( 38.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eWeight_Change (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eLose\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e186 ( 2.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e169 ( \u0026nbsp;2.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e17 ( \u0026nbsp;3.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.453\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo Change\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e6402 (97.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e5921 ( 97.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e481 ( 96.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eIncrease\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e11 ( 0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e11 ( \u0026nbsp;0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e0 ( \u0026nbsp;0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eSBP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e131 [118, 147]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e132 [119, 147]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e130 [115, 146]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eDBP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e79 [70, 88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e79 [70, 88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e80 [70, 90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003ePR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e78 [70, 88]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e78 [69, 87]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e80 [70, 92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTemperature\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e36.50 [36.30, 36.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e36.50 [36.30, 36.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e36.50 [36.50, 36.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eRR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e20 [19, 20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e20 [19, 20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e20 [20, 21]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eHypertension (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3923 (61.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e3028 ( 62.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e895 ( 57.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e2435 (38.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1783 ( 37.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e652 ( 42.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eDiabetes (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e6291 (88.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4827 ( 89.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1464 ( 86.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e782 (11.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e551 ( 10.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e231 ( 13.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eGGT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e25.00 [17.00, 45.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e24.80 [17.00, 43.48]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e27.00 [17.00, 50.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNeut\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e4.78 [3.47, 7.01]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4.64 [3.41, 6.66]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e5.60 [3.80, 8.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eLDL-C\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e2.49 [1.93, 3.13]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e2.50 [1.94, 3.15]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e2.45 [1.92, 3.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eHDL-C\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e1.19 [0.99, 1.45]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1.19 [0.99, 1.45]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1.19 [0.98, 1.42]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.132\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eMono\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e0.37 [0.27, 0.52]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e0.35 [0.26, 0.49]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e0.46 [0.32, 0.66]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eBasophil\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e0.02 [0.01, 0.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e0.02 [0.01, 0.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e0.02 [0.01, 0.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eEos\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e0.09 [0.04, 0.17]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e0.09 [0.04, 0.17]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e0.08 [0.03, 0.15]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eAST\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e24.00 [18.82, 39.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e23.50 [18.59, 34.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e29.00 [19.00, 88.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eUrea\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e5.66 [4.58, 7.06]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e5.60 [4.54, 6.94]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e5.90 [4.60, 7.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eUric Acid\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e328.80 [268.00, 400.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e327.75 [267.60, 398.08]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e332.00 [269.00, 406.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.059\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eMCV\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e92.00 [88.60, 95.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e91.70 [88.30, 95.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e93.00 [90.00, 97.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eMCHC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e331 [322, 339]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e332 [324, 340]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e325 [317, 333]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eMCH\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e30.50 [29.40, 31.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e30.50 [29.40, 31.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e30.30 [29.20, 31.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTBIL\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e10.60 [7.60, 14.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e11.00 [8.20, 14.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e8.90 [5.00, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e67.99 [63.30, 72.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e68.90 [64.10, 73.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e64.60 [60.50, 69.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTG\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e1.31 [0.95, 1.92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1.31 [0.94, 1.92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1.31 [0.97, 1.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.732\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eWBC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e7.00 [5.50, 9.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e6.83 [5.43, 8.86]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e7.80 [5.80, 10.30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eALB\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e41.00 [37.80, 44.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e41.50 [38.40, 44.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e39.00 [36.20, 42.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eDBIL\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3.50 [2.40, 5.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e3.60 [2.50, 5.06]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e3.10 [2.20, 4.60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\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: 16.6667%;\"\u003e\n \u003cp\u003eALP\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e75 [61, 92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e74.90 [61, 92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e75 [61, 92]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.579\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eRDW-SD\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e43.30 [40.10, 46.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e44.00 [41.60, 47.10]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e12.70 [12.20, 13.40]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\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: 16.6667%;\"\u003e\n \u003cp\u003eRDW-CV\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e13.20 [12.70, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e13.20 [12.70, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e12.90 [12.30, 13.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\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: 16.6667%;\"\u003e\n \u003cp\u003eCreatinine\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e69.40 [58.00, 84.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e69.25 [57.90, 83.70]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e70.00 [58.00, 85.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.029\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eHb\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e133 [121, 145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e133 [122, 145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e133 [120, 145]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.156\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003ePLT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e190.00 [153.00, 233.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e190.00 [153.00, 231.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e193.00 [154.25, 243.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eTC\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e4.41 [3.70, 5.21]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e4.43 [3.69, 5.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e4.35 [3.70, 5.13]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e0.08\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eALT\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e22.00 [14.64, 35.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e21.00 [14.20, 32.80]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e25.00 [16.00, 42.50]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eK+\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e3.95 [3.70, 4.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e3.98 [3.70, 4.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e3.90 [3.60, 4.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eNa+\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e140.90 [138.60, 143.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e141.00 [138.80, 143.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e140.00 [138.00, 142.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 16.6667%;\"\u003e\n \u003cp\u003eLymph\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 24.2574%;\"\u003e\n \u003cp\u003e1.40 [1.01, 1.84]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 22.7723%;\"\u003e\n \u003cp\u003e1.43 [1.04, 1.87]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 25.7426%;\"\u003e\n \u003cp\u003e1.31 [0.95, 1.73]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10.5611%;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch3\u003e3.2.\u0026nbsp;Risk factors\u003c/h3\u003e\n\u003cp\u003eThe highest performance point was\u0026nbsp;40\u0026nbsp;as estimated by CatBoost-based RFE with 5-fold cross-validation (CV) (Fig.S2).\u0026nbsp;However, upon further examination, we selected 21 variables as the optimal point. This decision was based on the fact that the model\u0026apos;s performance at 21 variables was nearly identical to that at\u0026nbsp;40\u0026nbsp;variables, offering a more parsimonious model with minimal loss in accuracy.\u003c/p\u003e\n\u003ch3\u003e3.3. Candidate algorithm screening\u003c/h3\u003e\n\u003cp\u003eThe results of 15 algorithms constructed using the scikit-learn package are presented in Table S2. Based on the model\u0026apos;s accuracy and AUC, the top-performing four algorithms (with accuracy \u0026gt; 0.753 and AUC \u0026gt; 0.805) selected for constructing the CAD prediction model for chest pain patients are Random Forest, CatBoost, XGBoost, and Gradient Boosting Classifier. In this study, the traditional Logistic Regression is chosen as the baseline for model comparison.\u003c/p\u003e\n\u003ch3\u003e3.4.\u0026nbsp;Prediction effects of different models\u003c/h3\u003e\n\u003cp\u003eThe corresponding ROC curves for the 5 models were shown in Fig.1. The Random Forest \u0026nbsp;model had the largest AUC of 0.829(95% CI: 0.810\u0026ndash;0.848). The AUCs of CatBoost, XGBoost, Gradient Boosting, and Logistic were 0.816 (95% CI: 0.794\u0026ndash;0.836), 0.820 (95% CI: 0.801\u0026ndash;0.839), 0.817 (95% CI: 0.800\u0026ndash;0.836), 0.736(95% CI: 0.711\u0026ndash;0.758), respectively. The accuracy, sensitivity, precision, and F1 scores of the 5 models were also calculated (Table 3) to comprehensively evaluate the performance of each model. The AUC of the XGBoost model (0.820 was second only to random forest (0.829), and it had the highest sensitivity (0.938) and the \u0026nbsp;F1 score (0.841) compared with the other 4 models.\u003c/p\u003e\n\u003cp\u003eTable 3.\u0026nbsp;Performance evaluation of five prediction models in internal validation.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"656\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModels\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrecision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eF1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandom Forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.829 (0.810\u0026ndash;0.848)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.771\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.859\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.821\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCatBoost\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.816 (0.794\u0026ndash;0.836)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.765\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.851\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.785\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.816\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXGBoost\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.820\u0026nbsp;(0.801\u0026ndash;0.839)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.766\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.938\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.746\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.831\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGradient Boosting\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.817 (0.800\u0026ndash;0.836)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.770\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.850\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.791\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.819\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLogistic Regression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.735 (0.711\u0026ndash;0.758)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.699\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.722\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.773\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.746\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch3\u003e3.5.External\u0026nbsp;Validation\u003c/h3\u003e\n\u003cp\u003eTo assess the generality and robustness of the CAD prediction model we developed for patients with chest pain, we performed external validation using a dataset of 1912 patients from another center.\u003c/p\u003e\n\u003cp\u003eOur results showed that although the predictive power of the four prediction models declined slightly in external validation, they maintained good discriminative power in the XGBoost model, with an ROC-AUC of 0.705, indicating good accuracy in distinguishing patients with and without CAD (Figure 2). In addition, the model shows good accuracy and calibration, suggesting accurate risk estimation(Table 4). These findings suggest that our machine learning-based prediction model for CAD retains its predictive ability in an external patient population, underscoring its potential utility in real-world clinical settings.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 4 presents the performance of five predictive models. XGBoost achieved the highest\u0026nbsp;Accuracy (0.763),\u0026nbsp;Sensitivity (0.931) and F1 score (0.845), indicating its superior ability to identify true positive cases and maintain a balance between Precision and Sensitivity. Although CatBoost showed the highest AUC (0.786) and strong Precision (0.809), its overall performance did not surpass XGBoost\u0026nbsp;when considering the critical balance between Sensitivity and Precision. The AUC of XGBoost was 0.763, which, although not the highest, complements its high Sensitivity and F1 score, reinforcing its reliability as a robust predictive model.\u003c/p\u003e\n\u003cp\u003eTable 4\u0026nbsp;Performance evaluation of five prediction models in external validation.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"656\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eModels\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrecision\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eF1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandom Forest\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.780\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.747\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.825\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.815\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.820\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGradient Boosting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.771\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.762\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.864\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.808\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.835\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXGBoost\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.763\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.763\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.931\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.774\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.845\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCatBoost\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.786\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.734\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.810\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.809\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.809\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLogistic Regression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.701\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.523\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.827\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.539\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch3\u003e3.6.\u0026nbsp;Model interpretation\u003c/h3\u003e\n\u003cp\u003eThe feature importance rankings of the 4 prediction models are shown in Fig.\u0026ensp;2, including \u0026nbsp;Random Forest (A),Gradient Boosting(B), XGBoost (C), and CatBoost (D). The importance scores were calculated using the built-in attributes in different ML algorithms. In these 4 models, the risk factors most associated with CAD in chest pain patients were Age;Mono; and HDL-C, AST, and ALT.\u003c/p\u003e\n\u003cp\u003eThis study illustrated how some characteristics influenced CAD in chest pain patients by incorporating the SHAP framework. Fig.\u0026ensp;3A shows the risk factors evaluated using the mean absolute SHAP value in the XGBoost prediction model. Fig.\u0026ensp;3B shows the features that impact the outcome. In each important feature row, different color dots represented the final impact of the feature on the outcome, where the red dots represented high-risk value and the blue represented low-risk value. Older age and higher Mono, RDW-SD, AST, and TG were associated with a higher predicted probability of CAD in chest pain patients.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe SHAP values indicate the prediction-related features of individual patients and the contribution of each feature to the prediction of prevalence. This study presented a sample of 2 individual predictions, with the red bars showing that the listed characteristics increase the risk of CAD in chest pain patients and the blue bars showing that the listed characteristics decrease the risk of CAD. Fig.\u0026ensp;3C shows a 60-year-old patient, whose Pulse Rate, Low-Density Lipoprotein Cholesterol, and Triglycerides were 81, 2.51 mmol/L, and 1.14 mmol/L, respectively. The SHAP value was -1.67. Fig.\u0026ensp;3D is a 89-year-old patient, whose Albumin, Pulse Rate, and High-Density Lipoprotein Cholesterol were 38.2 g/L, 55 , and 0.73 mmol/L, respectively. The SHAP value was 6.42.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eIn this study, we constructed and validated an interpretable ML-based prediction model to predict CAD in chest pain patients. Among the five ML prediction models, the XGBoost model showed the best performance during internal validation, with an AUC of 0.820 (0.801–0.839), Sensitivity of 0.938, and an F1 score of 0.831, surpassing the performance of the other four models.\u0026nbsp;Subsequently, during external validation, the XGBoost model exhibited commendable performance with an AUC of 0.763, Sensitivity of 0.931, and an F1 score of 0.845, highlighting its robustness when applied to an independent dataset. While there was a general decline in performance metrics during external validation, which is expected due to the challenges of generalizing to new data, XGBoost retained its predictive power and balance between Precision and Sensitivity better than the other models. This consistent performance across both internal and external validations underscores XGBoost's reliability and effectiveness as the best choice for predicting coronary artery disease in chest pain patients.\u0026nbsp;In recent studies, XGBoost\u0026nbsp;has been commonly used to build predictive models and has shown excellent discriminative power in many studies.\u0026nbsp;Ding, LanPing\u0026nbsp;et\u0026nbsp;al.\u003csup\u003e[31]\u003c/sup\u003e developed A ML-based approach, showed optimum performance and might help predict HTPR on clopidogrel after PCI and guide clinical decision-making using\u0026nbsp;9\u0026nbsp;ML algorithms, and the\u0026nbsp;XGBoost\u0026nbsp;showed the best performance, with an AUC of 0.82, a precision of 0.80, a recall of 0.44, an F1 score of 0.57, and an accuracy of 0.87. Meanwhile, the importance of features based on SHAP values in this study demonstrated that the ML approach can explain key features of\u0026nbsp;CAD in chest pain patients,\u0026nbsp;as well as that visualization of SHAP summary plots and force maps of the\u0026nbsp;XGBoost model\u0026nbsp;can allow clinicians to visualize and understand key features.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn general, the contributions of this study are as follows: first, we utilized a comprehensive set of 15 machine learning models and subsequently selected the most effective\u0026nbsp;four\u0026nbsp;for external validation. Other advanced ML knowledge was used in this study, such as missing value filling based on KNN, feature selection based on RFECV, and the random undersampling with SMOTE oversampling technique to address sample imbalance. The results showed that these methods can effectively improve the prediction of the risk of\u0026nbsp;CAD in chest pain patients.\u003c/p\u003e\n\u003cp\u003eSecond, it is always challenging to correctly interpret the large arrays of clinical data of predictive models constructed based on ML and visually present the predicted results to clinicians. Therefore, this study applied the SHAP framework to the\u0026nbsp;XGBoost model\u0026nbsp;“black-box” tree integration model to achieve optimal prediction and interpretability. It also helps clinicians better understand the decision-making process of prediction models and facilitates the use of prediction results, instead of blindly trusting the results of algorithms. Furthermore, the predictive model both took into account the key risk factors and visually explained to clinicians the characteristics that were likely to contribute to a higher (or lower) risk of\u0026nbsp;CAD in chest pain patients. As shown in\u0026nbsp;\u003ca href=\"http://www-sciencedirect-com-s.swebvpn.cqmu.edu.cn:8118/science/article/pii/S0939475323002338?via=ihub#fig4\"\u003eFig. 3\u003c/a\u003eB, the higher the concentration of\u0026nbsp;age,\u0026nbsp;monocyte\u0026nbsp;count,\u0026nbsp;Aspartate Aminotransferase,\u0026nbsp;and triglyceride level\u0026nbsp;in chest pain patients, the greater the risk of\u0026nbsp;CAD.\u0026nbsp;Aspartate Aminotransferase\u0026nbsp;(AST) is an enzyme primarily found in the liver and heart, playing a crucial role in amino acid metabolism. AST levels are independently positively associated with the risk and severity of premature CAD, suggesting that these enzymes could serve as surrogate markers for cardiovascular risk in this specific group of patients\u003csup\u003e[32]\u003c/sup\u003e. Recent studies have also indicated\u0026nbsp;that\u0026nbsp;peripheral monocyte count above 0.45 k/uL may be considered as a predictor of significant CAD in symptomatic patients with chronic coronary syndrome\u003csup\u003e[33]\u003c/sup\u003e. A higher\u0026nbsp;triglyceride level\u0026nbsp;index may be independently associated with a higher incidence of ASCVDs, CAD, and stroke in people without ASCVDs at baseline\u003csup\u003e[34]\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlthough this study has several strengths, it also suffers from several limitations. Firstly, the observed decrease in the model's performance during external validation could potentially be attributed to the utilization of data from multiple hospitals, which introduces inherent variations in data collection protocols across different institutions. These variations may have led to disparities in the distribution and characteristics of data between the internal and external validation datasets, thereby impacting the model's ability to generalize effectively beyond the training dataset. As information technology advances and policies progress, data standardization across institutions is expected to improve, which would enhance the model's generalizability and reproducibility in real-world healthcare applications. Despite these challenges, the insights gained from internal validation underscore the potential clinical utility of the model and highlight the importance of ongoing efforts to refine predictive models for accurate and reliable clinical decision-making.\u003c/p\u003e\n\u003cp\u003eSecondly, the absence of image data, particularly key echocardiographic parameters unavailable from the Chongqing Medical University Medical Data Platform, represents a limitation. Future work could explore incorporating multimodal data for a more comprehensive analysis.\u0026nbsp;Additionally, the reliance on laboratory test indicators and comorbidities as the primary variables in our study, although essential for ensuring the prediction model's performance and identifying risk factors, resulted in\u0026nbsp;21\u0026nbsp;features. This abundance of features may pose challenges in the practical application of the model within a clinical setting.\u0026nbsp;Following\u0026nbsp;this, we further recommend integrating the model into hospital information systems\u0026nbsp;rather than requiring healthcare personnel to independently utilize it. By seamlessly integrating the model into existing HIS infrastructure, healthcare providers can access predictive insights directly within their workflow, without the need for additional training or expertise in data analysis.\u003c/p\u003e\n\u003cp\u003eMoreover, despite these limitations, our model holds promise for various practical applications. For instance, it could serve as a valuable tool for coronary artery disease (CAD) screening in primary care settings and aid in auxiliary diagnosis. Leveraging clinical data from patients, the model can efficiently identify high-risk individuals, offer decision support for clinicians in selecting appropriate diagnostic procedures, and guide the formulation of personalized treatment plans. Particularly in resource-constrained medical environments, the application of our model can assist healthcare providers in optimizing the use of limited resources, thereby improving early CAD detection rates and treatment outcomes. Although further research and validation are warranted, we remain optimistic about the potential clinical utility of our model and anticipate that future developments will validate its effectiveness and reliability in real-world clinical settings.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eIn this study, we successfully developed an advanced XGBoosting model for predicting the risk of CAD in individuals experiencing chest pain. Our model demonstrated superior performance compared with traditional approaches. Notably, age, monocyte count, Aspartate Aminotransferase, and triglyceride level emerged as significant risk factors associated with CAD in chest pain patients. The identification of these key factors not only enhances our understanding of the contributors to CAD but also provides valuable insights for more targeted interventions and risk management strategies in clinical settings. As we conclude, this research serves as a foundational step, laying the groundwork for future investigations that seek to further enhance the precision and efficacy of predicting early CAD risk in individuals presenting with chest pain. The findings presented here underscore the potential for continuous advancements in predictive modeling, ultimately contributing to improved patient outcomes and informed clinical decision-making.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDeclaration of generative AI and AI-assisted technologies in the writing process\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring the preparation of this work the author(s) used ChatGTP3.5 in order to improve readability and language. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe present study complied with the principles of the Declaration of Helsinki and was approved by The Ethics Committee of ChongQing Medical University (Ethics Number: 2023044). Given the retrospective nature of the study on an anonymized database, informed consent was waived by The Ethics Committee of ChongQing Medical University.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets for this study can be found in Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University (https://demo.yiducloud.com.cn).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was Sponsored by Natural Science Foundation of Chongqing (CSTB2022NSCQ-MSX0837) and the Intelligent Medicine Research Project of Chongqing Medical University (YJSZHYX202212).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors thank their respective institutions for their support. We are also thankful to the Natural Science Foundation of Chongqing and the Intelligent Medicine Research Project of Chongqing Medical University for funding this project .\u003c/p\u003e\n\u003cp\u003eThe datasets for this study can be found in Medical Big Data Platform of the Medical Data Research Institute of Chongqing Medical University (https://demo.yiducloud.com.cn).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePagliaro B R, Cannata F, Stefanini G G and Bolognese L (2020) Myocardial ischemia and coronary disease in heart failure\u003cem\u003e \u003c/em\u003eHEART FAIL REV 25(1): 53-65\u003c/li\u003e\n\u003cli\u003eMedina-Leyte D J, Zepeda-Garcia O, Dominguez-Perez M, Gonzalez-Garrido A, Villarreal-Molina T and Jacobo-Albavera L (2021) Endothelial Dysfunction, Inflammation and Coronary Artery Disease: Potential Biomarkers and Promising Therapeutical Approaches\u003cem\u003e \u003c/em\u003eINT J MOL SCI 22(8):\u003c/li\u003e\n\u003cli\u003eShreya D, Zamora D I, Patel G S, Grossmann I, Rodriguez K, Soni M, Joshi P K, Patel S C and Sange I (2021) Coronary Artery Calcium Score - A Reliable Indicator of Coronary Artery Disease?\u003cem\u003e \u003c/em\u003eCureus 13(12): e20149\u003c/li\u003e\n\u003cli\u003eRalapanawa U and Sivakanesan R (2021) Epidemiology and the Magnitude of Coronary Artery Disease and Acute Coronary Syndrome: A Narrative Review\u003cem\u003e \u003c/em\u003eJOURNAL OF EPIDEMIOLOGY AND GLOBAL HEALTH 11(2): 169-77\u003c/li\u003e\n\u003cli\u003eRoth G A, Abate D, Abate K H, Abay S M, Abbafati C, Abbasi N, Abbastabar H,... GBD C D C (2018) Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017\u003cem\u003e \u003c/em\u003eLANCET 392(10159): 1736-88\u003c/li\u003e\n\u003cli\u003eBeltrame J F, Crea F, Kaski J C, Ogawa H, Ong P, Sechtem U, Shimokawa H, Merz C N B and Coronary V D I (2017) International standardization of diagnostic criteria for vasospastic angina\u003cem\u003e \u003c/em\u003eEUR HEART J 38(33): 2565-8\u003c/li\u003e\n\u003cli\u003eKnuuti J, Wijns W, Saraste A, Capodanno D, Barbato E, Funck-Brentano C, Prescott E,... European S C (2020) 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes The Task Force for the diagnosis and management of chronic coronary syndromes of the European Society of Cardiology (ESC)\u003cem\u003e \u003c/em\u003eEUR HEART J 41(3): 407-77\u003c/li\u003e\n\u003cli\u003eTamis-Holland J E, Jneid H, Reynolds H R, Agewall S, Brilakis E S, Brown T M, Lerman A,... Council Q C O R (2019) Contemporary Diagnosis and Management of Patients With Myocardial Infarction in the Absence of Obstructive Coronary Artery Disease: A Scientific Statement From the American Heart Association\u003cem\u003e \u003c/em\u003eCIRCULATION 139(18): E891-908\u003c/li\u003e\n\u003cli\u003eGenders T, Steyerberg E W, Alkadhi H, Leschka S, Desbiolles L, Nieman K, Galema T W,... CAD C (2011) A clinical prediction rule for the diagnosis of coronary artery disease: validation, updating, and extension\u003cem\u003e \u003c/em\u003eEUR HEART J 32(11): 1316-30\u003c/li\u003e\n\u003cli\u003eDIAMOND G A and FORRESTER J S (1979) ANALYSIS OF PROBABILITY AS AN AID IN THE CLINICAL-DIAGNOSIS OF CORONARY-ARTERY DISEASE\u003cem\u003e \u003c/em\u003eNEW ENGL J MED 300(24): 1350-8\u003c/li\u003e\n\u003cli\u003eKnuuti J, Wijns W, Saraste A, Capodanno D, Barbato E, Funck-Brentano C, Prescott E,... European S C (2020) 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes The Task Force for the diagnosis and management of chronic coronary syndromes of the European Society of Cardiology (ESC)\u003cem\u003e \u003c/em\u003eEUR HEART J 41(3): 407-77\u003c/li\u003e\n\u003cli\u003eFihn S D, Gardin J M, Abrams J, Berra K, Blankenship J C, Dallas A P, Douglas P S,... Yancy C W (2012) 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS Guideline for the Diagnosis and Management of Patients With Stable Ischemic Heart Disease\u003cem\u003e \u003c/em\u003eJ AM COLL CARDIOL 60(24): E44-164\u003c/li\u003e\n\u003cli\u003eSkinner J S, Smeeth L, Kendall J M, Adams P C, Timmis A and Chest P G D G (2010) NICE guidance. Chest pain of recent onset: assessment and diagnosis of recent onset chest pain or discomfort of suspected cardiac origin\u003cem\u003e \u003c/em\u003eHEART 96(12): 974-8\u003c/li\u003e\n\u003cli\u003eGoldstein B A, Navar A M and Carter R E (2017) Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges\u003cem\u003e \u003c/em\u003eEUR HEART J 38(23): 1805-14\u003c/li\u003e\n\u003cli\u003eAli M M, Gul S, Naqvi M, Hakam L, Inayat A, Saleem S, Polavarpu M and Syed M A (2021) Utility of Coronary Artery Calcium Scores in Predicting Risk of Subclinical Cardiovascular Atherosclerotic Disease: An Analysis of Limitations to its Adoption With Policy Recommendations.\u003cem\u003e \u003c/em\u003eCureus 13(4): e14647\u003c/li\u003e\n\u003cli\u003eDoolub G, Mamalakis M, Alabed S, Van der Geest R J, Swift A J, Rodrigues J C L, Garg P, Joshi N V and Dastidar A (2023) Artificial Intelligence as a Diagnostic Tool in Non-Invasive Imaging in the Assessment of Coronary Artery Disease.\u003cem\u003e \u003c/em\u003eMedical sciences (Basel, Switzerland) 11(1):\u003c/li\u003e\n\u003cli\u003eXu Z, Pan J, Chen T, Zhou Q, Wang Q, Cao H, Fan F,... Wang D (2018) A prediction score for significant coronary artery disease in Chinese patients \u0026ge;50 years old referred for rheumatic valvular heart disease surgery\u003cem\u003e \u003c/em\u003eINTERACT CARDIOV TH 26(4): 623-30\u003c/li\u003e\n\u003cli\u003eAndrikou I, Tsioufis C, Dimitriadis K, Konstantinidis D, Kasiakogias A, Kouremeti M, Andrikou E,... Tousoulis D (2018) Uric acid as an independent predictor of coronary artery disease in essential hypertension: Data from an 8-year-follow-up study\u003cem\u003e \u003c/em\u003eCLIN EXP PHARMACOL P 45(8): 866-9\u003c/li\u003e\n\u003cli\u003eCheng Y, Chen Y, Mao M, Wang R, Zhu J and He Q (2023) Association of inflammatory indicators with intensive care unit mortality in critically ill patients with coronary heart disease\u003cem\u003e \u003c/em\u003eFRONT IMMUNOL 14(\u003c/li\u003e\n\u003cli\u003eWang C, Zhao Y, Jin B Y, Gan X D, Liang B, Xiang Y, Zhang X K, Lu Z B and Zheng F (2021) Development and Validation of a Predictive Model for Coronary Artery Disease Using Machine Learning\u003cem\u003e \u003c/em\u003eFRONTIERS IN CARDIOVASCULAR MEDICINE 8(\u003c/li\u003e\n\u003cli\u003eGaravand A, Salehnasab C, Behmanesh A, Aslani N, Zadeh A H and Ghaderzadeh M (2022) Efficient Model for Coronary Artery Disease Diagnosis: A Comparative Study of Several Machine Learning Algorithms\u003cem\u003e \u003c/em\u003eJ HEALTHC ENG 2022(\u003c/li\u003e\n\u003cli\u003eLey C, Martin R K, Pareek A, Groll A, Seil R and Tischer T (2022) Machine learning and conventional statistics: making sense of the differences\u003cem\u003e \u003c/em\u003eKNEE SURG SPORT TR A 30(3): 753-7\u003c/li\u003e\n\u003cli\u003eOgami C, Tsuji Y, Seki H, Kawano H, To H, Matsumoto Y and Hosono H (2021) An artificial neural network-pharmacokinetic model and its interpretation using Shapley additive explanations\u003cem\u003e \u003c/em\u003eCPT-PHARMACOMETRICS \u0026amp; SYSTEMS PHARMACOLOGY 10(7): 760-8\u003c/li\u003e\n\u003cli\u003eLiao S G, Lin Y, Kang D D, Chandra D, Bon J, Kaminski N, Sciurba F C and Tseng G C (2014) Missing value imputation in high-dimensional phenomic data: imputable or not, and how?\u003cem\u003e \u003c/em\u003eBMC BIOINFORMATICS 15(\u003c/li\u003e\n\u003cli\u003eChen R, Dewi C, Huang S and Caraka R E (2020) Selecting critical features for data classification based on machine learning methods\u003cem\u003e \u003c/em\u003eJOURNAL OF BIG DATA 7(1):\u003c/li\u003e\n\u003cli\u003ePedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M,... Duchesnay E (2011) Scikit-learn: Machine Learning in Python\u003cem\u003e \u003c/em\u003eJ MACH LEARN RES 12(2825-30\u003c/li\u003e\n\u003cli\u003eAli M (2020) PyCaret: An open source, low-code machine learning library in Python. \u003c/li\u003e\n\u003cli\u003eMuqeet M, Malik H, Panhwar S, Khan I U, Hussain F, Asghar Z, Khatri Z and Mahar R B (2023) Enhanced cellulose nanofiber mechanical stability through ionic crosslinking and interpretation of adsorption data using machine learning\u003cem\u003e \u003c/em\u003eINT J BIOL MACROMOL 237(\u003c/li\u003e\n\u003cli\u003eFernandez A, Garcia S, Herrera F and Chawla N V (2018) SMOTE for Learning from Imbalanced Data: Progress and Challenges, Marking the 15-year Anniversary\u003cem\u003e \u003c/em\u003eJ ARTIF INTELL RES 61(863-905\u003c/li\u003e\n\u003cli\u003eRodriguez-Perez R and Bajorath J (2020) Interpretation of machine learning models using shapley values: application to compound potency and multi-target activity predictions\u003cem\u003e \u003c/em\u003eJ COMPUT AID MOL DES 34(10): 1013-26\u003c/li\u003e\n\u003cli\u003eDing L P, Li P, Yang L R, Pan M M, Zhou M, Zhang C, Yan Y D,... Gu Z C (2024) A novel machine learning model to predict high on-treatment platelet reactivity on clopidogrel in Asian patients after percutaneous coronary intervention\u003cem\u003e \u003c/em\u003eINT J CLIN PHARM-NET 46(1): 90-100\u003c/li\u003e\n\u003cli\u003eMasoudkabir F, Karbalai S, Vasheghani-Farahani A, Aliabadi L L, Boroumand M A, Aiatollahzade-Esfahani F, Pashing M,... Saadat S (2011) The Association of Liver Transaminase Activity With Presence and Severity of Premature Coronary Artery Disease\u003cem\u003e \u003c/em\u003eANGIOLOGY 62(8): 614-9\u003c/li\u003e\n\u003cli\u003eUrbanowicz T, Olasinska-Wisniewska A, Michalak M, Komosa A, Filipiak K J, Uruski P, Radziemski A, Tykarski A and Jemielity M (2023) Predictive role of monocyte count for significant coronary artery disease identification in patients with stable coronary artery disease\u003cem\u003e \u003c/em\u003eCARDIOL J \u003c/li\u003e\n\u003cli\u003eDing X B, Wang X Z, Wu J, Zhang M L and Cui M Z (2021) Triglyceride-glucose index and the incidence of atherosclerotic cardiovascular diseases: a meta-analysis of cohort studies\u003cem\u003e \u003c/em\u003eCARDIOVASC DIABETOL 20(1):\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Coronary artery disease, Chest pain, prediction model, machine learning, Multicenter study","lastPublishedDoi":"10.21203/rs.3.rs-5234204/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5234204/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCoronary artery disease (CAD) is a prevalent condition among chest pain patients, and accurate prediction of the disease is crucial to ensure timely interventions and improve patient outcomes. We aim to elaborate a prediction model for CAD in chest pain patients using machine learning approaches. A retrospective analysis was performed using electronic health records of patients who presented with chest pain at seven hospitals. A total of 8474 patients were included in the study, where 63.25% were diagnosed with CAD. The data included demographic information, medical history, and laboratory results. Machine learning algorithms, including Random Forest, CatBoost, XGBoosting, Gradient Boosting, Light Gradient, AdaBoost, Ridge Classifier, Linear Discriminant, Logistic Regression, Decision Tree, SVM, Quadratic Discriminant, K Neighbors, Naive Bayes, and Dummy Classifier were trained and evaluated to predict the presence of CAD.The prediction model achieved an overall accuracy of 0.766 in identifying CAD in chest pain patients. The sensitivity and precision were 0.938 and 0.746, respectively. Important predictors for CAD included age, pulse rate, monocyte, and red cell distribution width SD. The eXtreme Gradient Boosting showed the best performance (area under the receiver operating characteristics, AUROC, 0.820, and 95% CI, 0.801\u0026ndash;0.839) Additionally, the model demonstrated robust performance in the validation group. This study successfully developed and validated a prediction model for CAD in chest pain patients using machine learning techniques. The model exhibited good predictive ability and could aid in the early identification of CAD in clinical practice, potentially leading to appropriate interventions and improved patient outcomes.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Prediction Model for Coronary Artery Disease in Chest Pain Patients: A Real-World Multicenter Study Based on Machine Learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-26 13:53:11","doi":"10.21203/rs.3.rs-5234204/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3f66914f-22a8-4026-94b4-2ab3139f0460","owner":[],"postedDate":"November 26th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":39900373,"name":"Health sciences/Diseases/Cardiovascular diseases"},{"id":39900374,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2025-11-07T10:54:02+00:00","versionOfRecord":[],"versionCreatedAt":"2024-11-26 13:53:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5234204","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5234204","identity":"rs-5234204","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}