Machine learning based prediction of kidney function deterioration in infective endocarditis

Machine learning based prediction of kidney function deterioration in infective endocarditis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine learning based prediction of kidney function deterioration in infective endocarditis Min Woo Kang, Yoonjin Kang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4385746/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 3 You are reading this latest preprint version Abstract Background: Acute kidney injury in infective endocarditis presents significant management challenges in intensive care unit (ICU). We explored the role of mean blood pressure(MBP) at the time of ICU admission predicting kidney function outcomes in endocarditis patients using deep learning model, Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE). Methods: This study utilized data from the Medical Information Mart for Intensive Care III database. Patients with infective endocarditis admitted to intensive care unit were included in this study. A machine learning model was developed to predict the kidney function deterioration. SHapley Additive exPlanations (SHAP) were used to understand how variables affect kidney function. Moreover, the GANITE model, a causal inference deep learning model, was used to determine the effect of blood pressure to kidney function. Results. A total of 484 patients were included in the analysis, among whom 85(17.6%) experienced kidney deterioration. Light gradient boosting machine, extreme gradient boosting, and the ensemble model showed area under the receiver operating characteristics of 0.790, 0.772, and 0.785, respectively, on the test data, all achieving an accuracy of 0.828. SHAP value plots revealed that higher blood pressure predicted a lower likelihood of kidney deterioration. Analysis using the GANITE model revealed that maintaining MBP≥65mmHg resulted in a decrease in the probability of kidney deterioration by 12.9%. Conclusions: In patients with infective endocarditis in ICU, the maintenance of MBP≥65mmHg prevented the future kidney function deterioration after ICU admission. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Endocarditis presents significant management challenges, especially when patients require admission to the intensive care unit (ICU). 1 Deterioration of kidney function is a particularly critical concern in these patients, as it significantly impacts outcomes and complicates treatment strategies. 1 – 3 The Medical Information Mart for Intensive Care III (MIMIC-III) database, a large, publicly available dataset containing de-identified health-related data for over forty thousand patients who stayed in ICU settings, provides a unique resource for analyzing the predictors of kidney function deterioration in this patient population. 4 This study aims to explore the role of mean blood pressure at the time of ICU admission as a potential factor influencing kidney function outcomes in endocarditis patients. Maintaining adequate mean blood pressure is essential for renal perfusion and function, and previous research has suggested its importance in patient outcomes, particularly in the surgical context for infective endocarditis. 5 By employing machine learning algorithms to analyze data, this research seeks to identify key variables that impact kidney function deterioration. Furthermore, to understand the causal relationship between mean blood pressure and kidney function deterioration, we plan to utilize a causal inference deep learning model, specifically Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE), reflecting the advanced methodologies required in this research area. Patients and Methods Study population This study utilized data from the MIMIC-III database, focusing on patients with ICU admission records. We selected individuals with diagnoses related to endocarditis at the time of ICU admission. The diagnoses included a range of endocarditis conditions such as Syphilitic, Gonococcal, Meningococcal, Candidal, Histoplasmosis, Coxsackie endocarditis, and various forms of bacterial and rheumatic endocarditis. The patient population was randomly divided into training and test datasets in a 7:3 ratio for machine learning (ML) and deep learning (DL) analysis. Variables for analysis The analysis included demographic data (age, sex), initial vital signs (systolic and diastolic blood pressure, and heart rate), arterial oxygen saturation (SpO2), and laboratory data (white blood cell count, hemoglobin, hematocrit, platelet count, initial and baseline creatinine, bicarbonate, sodium, potassium, bacteremia types). Laboratory data selection was based on measurements taken closest to ICU admission, within a 24-hour pre-admission to 6-hour post-admission window. Surgical procedures and vital signs, including norepinephrine rate (mcg/kg/min) and intubation status within 6 hours of ICU admission, were also considered. Vancomycin treatment at the time of ICU admission was included as a variable. Kidney deterioration was defined as undergoing renal replacement therapy or a doubling of the initial creatinine level during the ICU stay. Continuous variables were compared using the t-test, while categorical variables were analyzed using the chi-square test. A multivariable logistic regression analysis was conducted to examine the association between kidney deterioration and various factors, with statistical significance set at P < 0.05. Machine learning and Deep learning Models predicting acute kidney injury The analysis was performed using Python's PyCaret package, evaluating several algorithms (Extra Trees (ET), CatBoost (CBT), Extreme Gradient Boosting (XGB), Random Forest (RF), Light Gradient Boosting Machine (LGBM), Gradient Boosting (GB), Decision Tree (DT), Ada Boost (AB), Logistic Regression (LR), Ridge, Linear Discriminant Analysis (LDA), Supporting Vector Machine (SVM)) through 10-fold cross-validation on the training data. The two algorithms with the highest area under the receiver operating characteristic (AUROC) were selected for further analysis, including variable importance assessment, hyperparameter tuning, and model performance evaluation on the test data. Model performance metrics included accuracy, AUROC, and F1-score. The models' receiver operating characteristic (ROC) curves and decision curve analysis (DCA) plots were compared to assess their effectiveness. Causal Inference Deep learning Model Following the ML model analysis, the dataset was divided into training and test sets for analysis using the GANITE model. This model treated mean blood pressure (MBP) ≥ 65 mmHg as the treatment and kidney deterioration as the outcome. Model performance was evaluated using accuracy and AUROC on the test dataset. The average treatment effect (ATE) of MBP ≥ 65 mmHg on kidney deterioration was calculated, along with the conditional average treatment effect (CATE) based on age, intubation within the first 6 hours, open heart surgery, bacteremia, initial creatinine > 2, and initial heart rate > 100. T-tests were used to compare CATEs. Additionally, the impact of MBP ≥ 65mmHg on the probability of kidney deterioration was analyzed, dividing the population into groups affected positively or negatively by the treatment and comparing variables between these groups using t-tests for continuous variables and chi-square tests for categorical variables. Additionally, SHapley Additive exPlanations (SHAP) values were examined. Results Patient Characteristics A total of 484 patients were included in the analysis, among whom 85 (17.6%) experienced kidney deterioration (Table 1 ). The average age of the patients was 59.25 years, and there was no statistically significant difference in age between those who did and did not experience kidney deterioration (p = 0.243). Males constituted 66% of the total cohort, with no significant difference in the proportion of kidney deterioration between genders (p = 0.274). In the group with kidney deterioration, diastolic blood pressure (DBP), hemoglobin, and hematocrit levels were significantly lower (p = 0.020, 0.03, 0.017, respectively), while white blood cell counts were significantly higher (p = 0.008). The data were randomly divided into 339 training data and 145 test data, with no significant differences in almost all variables between the two groups (Supplementary Table 1). Table 1 Baseline characteristics Total (N = 484) Kidney deterioration (N = 85) No Kidney deterioration (N = 399) P-value Age 59.25 (57.76–60.75) 61.08 (57.72–64.45) 58.86 (57.2-60.53) 0.243 Male 318 (0.66) 51 (0.6) 267 (0.67) 0.274 Systolic blood pressure (mmHg) 117.18 (115.08-119.27) 114.55 (108.49-120.62) 117.73 (115.53-119.93) 0.329 Diastolic blood pressure (mmHg) 60.68 (59.22–62.14) 56.85 (53.29–60.41) 61.5 (59.9-63.09) 0.02 Heart Rate (beats per minute) 93.84 (92.04–95.64) 94.21 (89.47–98.95) 93.76 (91.81–95.71) 0.861 SpO 2 (%) 96.92 (96.45–97.38) 97.13 (96.26-98.0) 96.87 (96.34–97.41) 0.619 WBC count *10 3 /㎕ 14.18 (13.49–14.87) 16.51 (14.54–18.47) 13.68 (12.96–14.4) 0.008 Hemoglobin (g/dL) 9.97 (9.8-10.14) 9.55 (9.12–9.97) 10.06 (9.88–10.24) 0.03 Hematocrit %(L/L) 30.05 (29.56–30.53) 28.69 (27.45–29.94) 30.33 (29.81–30.86) 0.017 Platelet count *10 3 /㎕ 238.17 (224.72-251.63) 217.67 (192.61-242.73) 242.54 (227.11-257.97) 0.096 Creatinine (mg/dL) 2.18 (1.97–2.4) 2.34 (1.74–2.95) 2.15 (1.92–2.38) 0.547 Base Creatinine (mg/dL) 1.9 (1.7–2.09) 1.54 (1.09–1.98) 1.97 (1.76–2.19) 0.082 Serum Na + (meq/l) 136.36 (135.88-136.84) 135.95 (134.85-137.05) 136.45 (135.91-136.98) 0.422 Serum K + (meq/l) 4.26 (4.2–4.33) 4.32 (4.15–4.48) 4.25 (4.18–4.33) 0.496 Open Heart Surgery 123 (0.25) 28 (0.33) 95 (0.24) 0.106 Bacteremia 136 (0.28) 28 (0.33) 108 (0.27) 0.337 MSSA 24 (0.05) 7 (0.08) 17 (0.04) 0.209 MRSA 30 (0.06) 7 (0.08) 23 (0.06) 0.542 Pseudomonas 2(< 0.01) 0 (0) 2(0.01) 0 Candidemia 2 (< 0.01) 0 (0) 2(0.01) 0 Vancomycin 37 (0.08) 8(0.09) 29 (0.07) 0.652 Bacterial Endocarditis 436 (0.9) 74 (0.87) 362 (0.91) 0.408 Candida Endocarditis 3 (0.01) 0 (0) 3(0.01) 0 Rheumatic Endocarditis 8 (0.02) 4 (0.05) 4(0.01) 0.05 Endocarditis NOS 36 (0.07) 7 (0.08) 29 (0.07) 0.936 Intubation 18 (0.04) 5 (0.06) 13 (0.03) 0.398 Norepinephrine (mcg/kg/min) 0.05 (0.02–0.07) 0.06 (0.0-0.12) 0.04 (0.02–0.07) 0.602 WBC = white blood cell; MSSA = Methicillin Sensitive Staphylococcus Aureus; MRSA = MSSA = Methicillin Resistant Staphylococcus Aureus; NOS = not otherwise specified Logistic regression for acute kidney injury A multivariable logistic regression analysis adjusted for all variables showed that higher age (Odds Ratio [OR] 1.02, p = 0.031), heart rate (OR 1.02, p = 0.011), initial creatinine (OR 8.99, p < 0.001), and white blood cell count (OR 1.05, p = 0.021) were associated with an increased risk of kidney deterioration. Conversely, lower diastolic blood pressure (OR 0.97, p = 0.021) was associated with a higher risk. Bacteremia (OR 1.80, p = 0.171) and vancomycin treatment (OR 2.30, p = 0.091) were not significantly correlated with the incidence of kidney deterioration (Table 2 ). Table 2 Odds ratio for kidney deterioration Odds Ratio (95% CI) P-value Male 0.76 (0.4–1.45) 0.408 Age 1.02 (1.0-1.04) 0.031 Systolic blood pressure 1.01 (0.99–1.02) 0.418 Systolic blood pressure 0.97 (0.95-1.0) 0.021 SpO 2 1.08 (0.98–1.19) 0.104 Heart rate 1.02 (1.01–1.04) 0.011 Initial Creatinine 8.99 (4.89–16.51) 0 Base creatinine 0.06 (0.02–0.13) 0 Initial bicarbonate 0.96 (0.9–1.03) 0.257 Initial WBC count 1.05 (1.01–1.08) 0.021 Initial hemoglobin 1.16 (0.59–2.29) 0.663 Initial platelet 1.0 (1.0–1.0) 0.098 Initial hematocrit 0.88 (0.7–1.11) 0.296 Initial sodium level 1.04 (0.98–1.11) 0.171 Initial Potassium level 1.06 (0.67–1.68) 0.793 Norepinephrine dose 4.87 (1.3-18.29) 0.019 Open heart surgery 1.69 (0.76–3.75) 0.196 Candida endocarditis 0.0 (0.0-inf) > 0.999 Bacterial endocarditis 100104412.13 (0.0-inf) > 0.999 Endocarditis NOS 275887203.26 (0.0-inf) > 0.999 Rheumatic endocarditis 268265741.5 (0.0-inf) > 0.999 Use of Vancomycin 2.3 (0.87–6.06) 0.091 Presence of bacteremia 1.8 (0.78–4.19) 0.171 MSSA 3.67 (0.93–14.5) 0.064 MRSA 1.29 (0.35–4.8) 0.701 Pseudomonas 0.0 (0.0-inf) > 0.999 Candidemia 0.0 (0.0-inf) 0.999 Intubation at admission 0.76 (0.15–3.88) 0.737 WBC = white blood cell; MSSA = Methicillin Sensitive Staphylococcus Aureus; MRSA = MSSA = Methicillin Resistant Staphylococcus Aureus; NOS = not otherwise specified Machine Learning Models predicting Kidney deterioration In a 10-fold cross-validation on the training data, LGBM and XGB models performed best, and hyperparameter tuning was conducted, followed by the development of an ensemble model of these two. LGBM, XGB, and the ensemble model showed AUROCs of 0.790, 0.772, and 0.785, respectively, on the test data, all achieving an accuracy of 0.828 (Table 3 and Fig. 1 ). The DCA plot indicated slightly better performance of XGB over LGBM in the test data (Fig. 2 ). Initial systolic blood pressure was the second most important variable in LGBM and seventh in XGB, whereas initial diastolic blood pressure was the eighth and fourth most important, respectively. SHAP value plots for both models' predictions on the test data showed a tendency of higher systolic and diastolic blood pressure predicting a lower likelihood of kidney deterioration. Other significant predictors included higher initial creatinine, lower hematocrit, higher initial heart rate, and older age. Table 3 Performance of models predicting kidney function deterioration Train AUC Train Accuracy Train F1-score Train Recall Test AUC Test Accuracy Test F1-score Test Recall Ensemble 0.922 0.947 0.83 0.759 0.785 0.828 0.194 0.111 LightGBM 0.918 0.944 0.819 0.741 0.790 0.828 0.194 0.111 XGBoost 0.925 0.947 0.833 0.776 0.772 0.828 0.242 0.148 LightGBM = Light Gradient Boosting Machine; XGBoost = Extreme Gradient Boosting; AUC = Area under the curve Causal Inference Deep Learning Model The GANITE model was trained to examine the Average Treatment Effect (ATE) of maintaining mean blood pressure (MBP) ≥ 65 mmHg, showing a decrease in the probability of kidney deterioration by 12.7% (95% Confidence Interval [CI]: -14.1 to -11.4%) in the training data, 13.4% (95% CI: -15.4 to -11.4%) in the test data, and 12.9% (95% CI: -14.0 to -11.8%) in the total data (Table 4 ). The GANITE model's AUROC was 0.716 in the test data, with an accuracy of 0.814. Patients undergoing open heart surgery showed a CATE of -15.3% for kidney deterioration at MBP ≥ 65mmHg, significantly larger than the − 12.1% CATE in patients not undergoing surgery (Table 5 ). Similarly, the CATE for patients over 60 years old was − 14.7%, significantly larger than the − 11.1% for those 60 years old or younger. Additionally, the effect of MBP ≥ 65 mmHg was significantly more pronounced in reducing the risk of kidney deterioration in patients with initial creatinine > 2.0 mg/dL and initial heart rate ≤ 100 beats per minutes. Groups with decreased vs. increased risk of kidney deterioration at MBP ≥ 65 showed significant differences in initial creatinine, baseline creatinine, and heart rate (Table 6 ). Table 4 Average treatment effect (ATE) and evaluation indexes of Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE) model Train Test Total Accuracy 0.838 0.814 0.831 AUROC 0.750 0.716 0.737 ATE -0.127 (-0.141–0.114) -0.134 (-0.154–0.114) -0.129 (-0.14–0.118) ATE = Average treatment effect; AUROC = area under the receiver operating characteristic; Table 5 conditional average treatment effect (CATE) for total population CATE for variable positive CATE for variable negative P-value Open heart surgery -0.153 (-0.17–0.135) -0.121 (-0.135–0.108) 0.005 Intubation -0.145 (-0.2–0.09) -0.128 (-0.14–0.117) 0.541 Bacteremia -0.129 (-0.153–0.106) -0.129 (-0.141–0.117) 0.977 Age > 60 -0.147 (-0.162–0.133) -0.111 (-0.127–0.094) 0.001 Initial Cr > 2mg/dL -0.147 (-0.167–0.127) -0.122 (-0.135–0.109) 0.038 Heart Rate > 100 -0.068 (-0.089–0.048) -0.162 (-0.174–0.151) 0 CATE = conditional average treatment effect;Cr = creatinine Table 6 Comparing negative effect group and positive effect group in total population ATE negative ATE positive P-value Open Heart Surgery 0.264 0.156 0.157 Male 0.663 0.6 0.496 Bacterial Endocarditis 0.902 0.889 0.984 Age 59.208 (57.633–60.782) 59.69 (54.814–64.566) 0.851 Creatinine 2.24 (2.004–2.476) 1.633 (1.304–1.963) 0.004 Heart rate 92.943 (91.093–94.793) 102.578 (95.826-109.329) 0.008 Base Creatinine 1.951 (1.74–2.162) 1.36 (1.065–1.655) 0.002 Hemoglobin 9.962 (9.788–10.136) 10.042 (9.37-10.714) 0.817 Serum Na + 136.392 (135.889-136.895) 136.067 (134.398-137.735) 0.709 ATE = Average treatment effect Discussion This research marks a significant advancement in predicting kidney function deterioration among ICU patients with endocarditis, employing both machine learning models and statistical techniques to understand the factors influencing kidney health. Notably, the study leverages a deep learning causal inference model to demonstrate the protective effect of maintaining a MBP ≥ 65 mmHg, underscoring its critical role in mitigating kidney deterioration risk under similar patient conditions. The study's findings, particularly the AUROC scores from the test dataset, highlight the predictive power of the LGBM model, underscoring the utility of machine learning in clinical settings. The use of SHAP values from tree-based machine learning models provides deeper insights into how various factors impact kidney deterioration. This approach offers a nuanced understanding beyond traditional logistic regression, which, while effective in illustrating associations between clinical variables and outcomes, may have limitations in assessing model fit and interpretability. The differential impact of SBP and DBP on kidney function, as revealed through logistic regression and machine learning analyses, emphasizes the complexity of blood pressure's role in kidney health. The study's progression to analyzing MBP through causal inference analysis addresses this complexity, providing a more holistic view of blood pressure's effects. The employment of a deep learning causal inference model represents a methodological strength, offering a way to approximate causality in scenarios where randomization is impractical. This aspect is particularly relevant in the ICU setting, where patient severity and ethical considerations limit the feasibility of randomized controlled trials. The study's ability to infer causality in such a constrained environment adds significant value to its findings, offering a model for future research in similar clinical contexts. Recent studies underscore the significant risk AKI poses to mortality in endocarditis patients. AKI is a frequent complication during IE, involving 69% of the patients. 6 Known risk factors of AKI is acute heart failure, usage of vancomycin and prosthetic valve involvement. 6 The specific role of antibiotics and their potential renal toxicity in the treatment of infective endocarditis have been documented, indicating the complexity of managing renal health in these patients. 7 , 8 AKI is associated with high mortality in IE (22.7% vs. 16.0%). 1 A nationwide study by Petersen et al. 9 revealed that dialysis-requiring AKI during admission for infective endocarditis is associated with a significantly increased one-year mortality risk from discharge. This finding highlights the critical nature of AKI as a high-risk factor for worse long-term outcomes compared to patients without dialysis requirements. 9 Ortiz-Soriano et al. 10 further support this by documenting the high morbidity and mortality associated with AKI in infective endocarditis patients, noting that two-thirds of such patients experience incident AKI, with those suffering from severe AKI incurring increased healthcare costs and a higher risk of mortality. 10 These studies collectively emphasize the detrimental impact of AKI on the survival of endocarditis patients, highlighting the urgent need for early detection and intervention to mitigate its effects. The significance of maintaining optimal blood pressure for kidney function, especially in septic conditions prevalent among endocarditis patients in the ICU, is well documented. A study by Jamme et al. 11 on amoxicillin crystalluria associated with AKI in patients treated for acute infective endocarditis illustrates the complex interplay between treatment modalities for endocarditis and kidney health. 11 This research emphasizes the importance of careful management of medications and supportive treatments to prevent further renal impairment in already at-risk populations. Septic shock and heart failure is well known risk factors of AKI in patients admitted for IE. 12 The decision to adopt a mean arterial pressure (MAP) target of ≥ 65 mmHg in the management of septic shock, as employed in our study, is firmly grounded in evidence-based guidelines and reinforced by the comprehensive literature review conducted by Leone et al. 13 . This analysis of seven comparative studies, encompassing both randomized clinical trials and observational studies, supports the adequacy of a 65 mmHg MAP target for most septic shock patients. This threshold is based on the premise that it generally suffices to sustain organ perfusion and function, thereby stabilizing the patient's hemodynamic status. The goal of the present study is not to presuppose conclusions but to rigorously investigate the potential protective role of mean blood pressure against kidney function deterioration in ICU-admitted endocarditis patients. Through this investigation, we aim to contribute valuable insights to the existing body of knowledge, facilitating the development of targeted interventions to improve patient outcomes. This approach emphasizes the critical need for comprehensive research into the interplay between renal health, treatment strategies, and patient management in endocarditis, aiming to enhance care and prognosis for this vulnerable patient population However, the study faces limitations, including its relatively small sample size of 484 subjects. While the specificity of the ICU setting and the condition of endocarditis inherently limit the availability of large datasets, this constraint may affect the generalizability of the findings. The absence of certain variables, such as vegetation size and ejection fraction, due to the limitations of the MIMIC-III database, further restricts the study's scope. These omissions highlight the need for more comprehensive datasets in future research to encompass a wider range of clinically relevant variables. Moreover, while the deep learning causal inference model provides valuable insights, it cannot fully substitute for randomization. This limitation underscores the inherent challenges in drawing causal inferences from observational data, especially in complex clinical settings like the ICU. Future studies might explore innovative methodologies to overcome these challenges, potentially incorporating hybrid models that blend machine learning with traditional epidemiological approaches to better approximate causal relationships. In conclusion, this study makes a compelling case for the importance of maintaining MBP ≥ 65 in ICU patients with endocarditis to reduce the risk of kidney function deterioration. Its methodological approach, combining statistical techniques with machine learning and deep learning models, offers a blueprint for future research aimed at unraveling the complex interplay of factors affecting patient outcomes in critical care settings. In conclusion, the machine learning model was effective in prediction of kidney deterioration in patients with infective endocarditis in ICU. Among these patients, the maintenance of MBP ≥ 65mmHg prevented the future kidney function deterioration after ICU admission. This study provides essential insights into kidney health management in critically ill patients, highlighting the potential of advanced analytics in enhancing patient care and outcomes. Declarations Acknowledgement: None Disclosure: No disclosure Conflict of interest: None Funding Declaration: None Ethics declaration: Not applicable Data Availability Declaration: not applicable (publicly available database) Conflict of interest: None Author Contribution MWK and YK wrote the manuscript and prepared all figures and tables. All authors reviewed the manuscript. References Bor DH, Woolhandler S, Nardin R, Brusch J, Himmelstein DU. Infective endocarditis in the U.S., 1998–2009: a nationwide study. PLoS One. 2013;8(3):e60033. Conlon PJ, Jefferies F, Krigman HR, Corey GR, Sexton DJ, Abramson MA. Predictors of prognosis and risk of acute renal failure in bacterial endocarditis. Clin Nephrol. 1998;49(2):96–101. Legrand M, Pirracchio R, Rosa A, Petersen ML, Van der Laan M, Fabiani JN, et al. Incidence, risk factors and prediction of post-operative acute kidney injury following cardiac surgery for active infective endocarditis: an observational study. Crit Care. 2013;17(5):R220. Johnson AE, Pollard TJ, Shen L, Lehman LW, Feng M, Ghassemi M, et al. MIMIC-III, a freely accessible critical care database. Sci Data. 2016;3:160035. Pericas JM, Hernandez-Meneses M, Munoz P, Martinez-Selles M, Alvarez-Uria A, de Alarcon A, et al. Characteristics and Outcome of Acute Heart Failure in Infective Endocarditis: Focus on Cardiogenic Shock. Clin Infect Dis. 2021;73(5):765–74. Gagneux-Brunon A, Pouvaret A, Maillard N, Berthelot P, Lutz MF, Cazorla C, et al. Acute kidney injury in infective endocarditis: A retrospective analysis. Med Mal Infect. 2019;49(7):527–33. Kunming P, Ying H, Chenqi X, Zhangzhang C, Xiaoqiang D, Xiaoyu L, et al. Vancomycin associated acute kidney injury in patients with infectious endocarditis: a large retrospective cohort study. Front Pharmacol. 2023;14:1260802. Barberan J, Mensa J, Artero A, Epelde F, Rodriguez JC, Ruiz-Morales J, et al. Factors associated with development of nephrotoxicity in patients treated with vancomycin versus daptomycin for severe Gram-positive infections: A practice-based study. Rev Esp Quimioter. 2019;32(1):22–30. Petersen JK, Jensen AD, Bruun NE, Kamper AL, Butt JH, Havers-Borgersen E, et al. Outcome of Dialysis-Requiring Acute Kidney Injury in Patients With Infective Endocarditis: A Nationwide Study. Clin Infect Dis. 2021;72(9):e232-e9. Ortiz-Soriano V, Donaldson K, Du G, Li Y, Lambert J, Cleland D, et al. Incidence and Cost of Acute Kidney Injury in Hospitalized Patients with Infective Endocarditis. J Clin Med. 2019;8(7). Jamme M, Oliver L, Ternacle J, Lepeule R, Moussafeur A, Haymann JP, et al. Amoxicillin crystalluria is associated with acute kidney injury in patients treated for acute infective endocarditis. Nephrol Dial Transplant. 2021;36(10):1955–8. Habib G, Lancellotti P, Antunes MJ, Bongiorni MG, Casalta JP, Del Zotti F, et al. 2015 ESC Guidelines for the management of infective endocarditis: The Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J. 2015;36(44):3075–128. Leone M, Asfar P, Radermacher P, Vincent JL, Martin C. Optimizing mean arterial pressure in septic shock: a critical reappraisal of the literature. Crit Care. 2015;19(1):101. Additional Declarations No competing interests reported. Supplementary Files SupplementaryTable1.docx Cite Share Download PDF Status: Under Review Version 1 posted Editor assigned by journal 08 May, 2024 Submission checks completed at journal 07 May, 2024 First submitted to journal 07 May, 2024 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-4385746","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":299988948,"identity":"ac630d6e-5974-469b-b098-aa10aff38eef","order_by":0,"name":"Min Woo Kang","email":"","orcid":"","institution":"Seoul National University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"Woo","lastName":"Kang","suffix":""},{"id":299988949,"identity":"b920f54f-3098-4405-9d54-8c4c9d1edf3c","order_by":1,"name":"Yoonjin Kang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIie3SvQrCMBDA8SsBXVq7RgTrI1wpFEEfxi526aCL6CYI9SEEn8OxEqhLZgnoUBdnHQsOnh9zyCiY/9KG3m/JFcBm+9U4AtLDqb5nZkwYmhPqNd3gRgQPq7LqT86Rv1kfF/VOgL8uWDTXEVmmyPEa87OcnTwpgMsRS6SOqCzmHMUQVDY+ObkAUMD2SxMSEJnWRAJTEqNKS/CIIJFER9qyHL9IFKqMdbw8dUOZrEIdadGNdfhDhFuVXu51Puh2D0K0daRX0Bo++3Dfm3TpL9ABgIA+O7f3a7PSTtpsNtv/9gTKk03Q4F1OagAAAABJRU5ErkJggg==","orcid":"","institution":"Seoul National University Hospital","correspondingAuthor":true,"prefix":"","firstName":"Yoonjin","middleName":"","lastName":"Kang","suffix":""}],"badges":[],"createdAt":"2024-05-08 01:27:00","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4385746/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4385746/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":56677000,"identity":"d2c6aed6-4da7-4dcd-ab1c-0f637c5a91d1","added_by":"auto","created_at":"2024-05-17 16:31:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":69849,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic curve of prediction models\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/a65283699b765818be33a826.png"},{"id":56677005,"identity":"6aba8ca7-5b1e-4dab-a168-2a3a38763f95","added_by":"auto","created_at":"2024-05-17 16:31:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":61863,"visible":true,"origin":"","legend":"\u003cp\u003eDecision curve analysis plots of prediction models\u003c/p\u003e","description":"","filename":"Figure21.png","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/25cc89ffc82cc81b6d4006f6.png"},{"id":56677001,"identity":"78cc8d09-e328-4a56-8ef0-ad08d21fd10a","added_by":"auto","created_at":"2024-05-17 16:31:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":90728,"visible":true,"origin":"","legend":"\u003cp\u003eVariable importance of Light Gradient Boosting Machine and Extreme Gradient Boosting models\u003c/p\u003e","description":"","filename":"Figure32.png","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/0fc3ad3df61fdf0b8f0582e0.png"},{"id":56677004,"identity":"c04c7c55-cc08-41cd-a4a0-3c1038cede39","added_by":"auto","created_at":"2024-05-17 16:31:40","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":136245,"visible":true,"origin":"","legend":"\u003cp\u003eSHapley Additive exPlanations of Light Gradient Boosting Machine and Extreme Gradient Boostingmodels\u003c/p\u003e","description":"","filename":"Figure41.png","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/7ae1fc7fbe741f6fd8157978.png"},{"id":56677002,"identity":"15b26dd4-318b-4d68-8c1c-f56d8eced079","added_by":"auto","created_at":"2024-05-17 16:31:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":39572,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristiccurve and Calibration plot of Generative Adversarial Nets for inference of Individualized Treatment Effects\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/81255dbbc928e206d0d56261.png"},{"id":56677024,"identity":"9cf215e3-8b58-4efc-a786-7bb69d452f6e","added_by":"auto","created_at":"2024-05-17 16:31:45","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1348069,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/8318017b-9e3c-430b-8fbb-d907101f66fc.pdf"},{"id":56677003,"identity":"61dc0249-b16f-4d41-b50d-8d58f99dd117","added_by":"auto","created_at":"2024-05-17 16:31:40","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":18199,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4385746/v1/25d24437c7fbc2a84518f03a.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine learning based prediction of kidney function deterioration in infective endocarditis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eEndocarditis presents significant management challenges, especially when patients require admission to the intensive care unit (ICU).\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e Deterioration of kidney function is a particularly critical concern in these patients, as it significantly impacts outcomes and complicates treatment strategies.\u003csup\u003e\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThe Medical Information Mart for Intensive Care III (MIMIC-III) database, a large, publicly available dataset containing de-identified health-related data for over forty thousand patients who stayed in ICU settings, provides a unique resource for analyzing the predictors of kidney function deterioration in this patient population.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThis study aims to explore the role of mean blood pressure at the time of ICU admission as a potential factor influencing kidney function outcomes in endocarditis patients. Maintaining adequate mean blood pressure is essential for renal perfusion and function, and previous research has suggested its importance in patient outcomes, particularly in the surgical context for infective endocarditis.\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eBy employing machine learning algorithms to analyze data, this research seeks to identify key variables that impact kidney function deterioration. Furthermore, to understand the causal relationship between mean blood pressure and kidney function deterioration, we plan to utilize a causal inference deep learning model, specifically Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE), reflecting the advanced methodologies required in this research area.\u003c/p\u003e"},{"header":"Patients and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy population\u003c/h2\u003e \u003cp\u003eThis study utilized data from the MIMIC-III database, focusing on patients with ICU admission records. We selected individuals with diagnoses related to endocarditis at the time of ICU admission. The diagnoses included a range of endocarditis conditions such as Syphilitic, Gonococcal, Meningococcal, Candidal, Histoplasmosis, Coxsackie endocarditis, and various forms of bacterial and rheumatic endocarditis. The patient population was randomly divided into training and test datasets in a 7:3 ratio for machine learning (ML) and deep learning (DL) analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eVariables for analysis\u003c/h2\u003e \u003cp\u003eThe analysis included demographic data (age, sex), initial vital signs (systolic and diastolic blood pressure, and heart rate), arterial oxygen saturation (SpO2), and laboratory data (white blood cell count, hemoglobin, hematocrit, platelet count, initial and baseline creatinine, bicarbonate, sodium, potassium, bacteremia types). Laboratory data selection was based on measurements taken closest to ICU admission, within a 24-hour pre-admission to 6-hour post-admission window. Surgical procedures and vital signs, including norepinephrine rate (mcg/kg/min) and intubation status within 6 hours of ICU admission, were also considered. Vancomycin treatment at the time of ICU admission was included as a variable. Kidney deterioration was defined as undergoing renal replacement therapy or a doubling of the initial creatinine level during the ICU stay. Continuous variables were compared using the t-test, while categorical variables were analyzed using the chi-square test. A multivariable logistic regression analysis was conducted to examine the association between kidney deterioration and various factors, with statistical significance set at P\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eMachine learning and Deep learning Models predicting acute kidney injury\u003c/h2\u003e \u003cp\u003eThe analysis was performed using Python's PyCaret package, evaluating several algorithms (Extra Trees (ET), CatBoost (CBT), Extreme Gradient Boosting (XGB), Random Forest (RF), Light Gradient Boosting Machine (LGBM), Gradient Boosting (GB), Decision Tree (DT), Ada Boost (AB), Logistic Regression (LR), Ridge, Linear Discriminant Analysis (LDA), Supporting Vector Machine (SVM)) through 10-fold cross-validation on the training data. The two algorithms with the highest area under the receiver operating characteristic (AUROC) were selected for further analysis, including variable importance assessment, hyperparameter tuning, and model performance evaluation on the test data. Model performance metrics included accuracy, AUROC, and F1-score. The models' receiver operating characteristic (ROC) curves and decision curve analysis (DCA) plots were compared to assess their effectiveness.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eCausal Inference Deep learning Model\u003c/h2\u003e \u003cp\u003eFollowing the ML model analysis, the dataset was divided into training and test sets for analysis using the GANITE model. This model treated mean blood pressure (MBP)\u0026thinsp;\u0026ge;\u0026thinsp;65 mmHg as the treatment and kidney deterioration as the outcome. Model performance was evaluated using accuracy and AUROC on the test dataset. The average treatment effect (ATE) of MBP\u0026thinsp;\u0026ge;\u0026thinsp;65 mmHg on kidney deterioration was calculated, along with the conditional average treatment effect (CATE) based on age, intubation within the first 6 hours, open heart surgery, bacteremia, initial creatinine\u0026thinsp;\u0026gt;\u0026thinsp;2, and initial heart rate\u0026thinsp;\u0026gt;\u0026thinsp;100. T-tests were used to compare CATEs. Additionally, the impact of MBP\u0026thinsp;\u0026ge;\u0026thinsp;65mmHg on the probability of kidney deterioration was analyzed, dividing the population into groups affected positively or negatively by the treatment and comparing variables between these groups using t-tests for continuous variables and chi-square tests for categorical variables. Additionally, SHapley Additive exPlanations (SHAP) values were examined.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePatient Characteristics\u003c/h2\u003e \u003cp\u003eA total of 484 patients were included in the analysis, among whom 85 (17.6%) experienced kidney deterioration (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The average age of the patients was 59.25 years, and there was no statistically significant difference in age between those who did and did not experience kidney deterioration (p\u0026thinsp;=\u0026thinsp;0.243). Males constituted 66% of the total cohort, with no significant difference in the proportion of kidney deterioration between genders (p\u0026thinsp;=\u0026thinsp;0.274). In the group with kidney deterioration, diastolic blood pressure (DBP), hemoglobin, and hematocrit levels were significantly lower (p\u0026thinsp;=\u0026thinsp;0.020, 0.03, 0.017, respectively), while white blood cell counts were significantly higher (p\u0026thinsp;=\u0026thinsp;0.008). The data were randomly divided into 339 training data and 145 test data, with no significant differences in almost all variables between the two groups (Supplementary Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal (N\u0026thinsp;=\u0026thinsp;484)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKidney deterioration (N\u0026thinsp;=\u0026thinsp;85)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo Kidney deterioration (N\u0026thinsp;=\u0026thinsp;399)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e59.25 (57.76\u0026ndash;60.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61.08 (57.72\u0026ndash;64.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e58.86 (57.2-60.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.243\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMale\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e318 (0.66)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51 (0.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e267 (0.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.274\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSystolic blood pressure (mmHg)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e117.18 (115.08-119.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e114.55 (108.49-120.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e117.73 (115.53-119.93)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.329\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDiastolic blood pressure (mmHg)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e60.68 (59.22\u0026ndash;62.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e56.85 (53.29\u0026ndash;60.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e61.5 (59.9-63.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHeart Rate (beats per minute)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e93.84 (92.04\u0026ndash;95.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e94.21 (89.47\u0026ndash;98.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e93.76 (91.81\u0026ndash;95.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.861\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSpO\u003c/b\u003e\u003csub\u003e\u003cb\u003e2\u003c/b\u003e\u003c/sub\u003e \u003cb\u003e(%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e96.92 (96.45\u0026ndash;97.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.13 (96.26-98.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e96.87 (96.34\u0026ndash;97.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.619\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eWBC count *10\u003c/b\u003e\u003csup\u003e\u003cb\u003e3\u003c/b\u003e\u003c/sup\u003e\u003cb\u003e/㎕\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14.18 (13.49\u0026ndash;14.87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.51 (14.54\u0026ndash;18.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.68 (12.96\u0026ndash;14.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHemoglobin (g/dL)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e9.97 (9.8-10.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9.55 (9.12\u0026ndash;9.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.06 (9.88\u0026ndash;10.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHematocrit %(L/L)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e30.05 (29.56\u0026ndash;30.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28.69 (27.45\u0026ndash;29.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30.33 (29.81\u0026ndash;30.86)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.017\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePlatelet count *10\u003c/b\u003e\u003csup\u003e\u003cb\u003e3\u003c/b\u003e\u003c/sup\u003e\u003cb\u003e/㎕\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e238.17 (224.72-251.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e217.67 (192.61-242.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e242.54 (227.11-257.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.096\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCreatinine (mg/dL)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.18 (1.97\u0026ndash;2.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.34 (1.74\u0026ndash;2.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.15 (1.92\u0026ndash;2.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.547\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBase Creatinine (mg/dL)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.9 (1.7\u0026ndash;2.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.54 (1.09\u0026ndash;1.98)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.97 (1.76\u0026ndash;2.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.082\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSerum Na\u003c/b\u003e\u003csup\u003e\u003cb\u003e+\u003c/b\u003e\u003c/sup\u003e \u003cb\u003e(meq/l)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e136.36 (135.88-136.84)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e135.95 (134.85-137.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e136.45 (135.91-136.98)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.422\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSerum K\u003c/b\u003e\u003csup\u003e\u003cb\u003e+\u003c/b\u003e\u003c/sup\u003e \u003cb\u003e(meq/l)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.26 (4.2\u0026ndash;4.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.32 (4.15\u0026ndash;4.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.25 (4.18\u0026ndash;4.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.496\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOpen Heart Surgery\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e123 (0.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28 (0.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e95 (0.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.106\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBacteremia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e136 (0.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e28 (0.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e108 (0.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.337\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMSSA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e24 (0.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (0.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.209\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMRSA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e30 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e23 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.542\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePseudomonas\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2(\u0026lt;\u0026thinsp;0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2(0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCandidemia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2 (\u0026lt;\u0026thinsp;0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2(0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVancomycin\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e37 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8(0.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.652\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBacterial Endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e436 (0.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e74 (0.87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e362 (0.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.408\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCandida Endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3 (0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3(0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRheumatic Endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e8 (0.02)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (0.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4(0.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEndocarditis NOS\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e36 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.936\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIntubation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e18 (0.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13 (0.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.398\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNorepinephrine (mcg/kg/min)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.05 (0.02\u0026ndash;0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.06 (0.0-0.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.04 (0.02\u0026ndash;0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.602\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eWBC\u0026thinsp;=\u0026thinsp;white blood cell; MSSA\u0026thinsp;=\u0026thinsp;Methicillin Sensitive Staphylococcus Aureus; MRSA\u0026thinsp;=\u0026thinsp;MSSA\u0026thinsp;=\u0026thinsp;Methicillin Resistant Staphylococcus Aureus; NOS\u0026thinsp;=\u0026thinsp;not otherwise specified\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eLogistic regression for acute kidney injury\u003c/h2\u003e \u003cp\u003eA multivariable logistic regression analysis adjusted for all variables showed that higher age (Odds Ratio [OR] 1.02, p\u0026thinsp;=\u0026thinsp;0.031), heart rate (OR 1.02, p\u0026thinsp;=\u0026thinsp;0.011), initial creatinine (OR 8.99, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and white blood cell count (OR 1.05, p\u0026thinsp;=\u0026thinsp;0.021) were associated with an increased risk of kidney deterioration. Conversely, lower diastolic blood pressure (OR 0.97, p\u0026thinsp;=\u0026thinsp;0.021) was associated with a higher risk. Bacteremia (OR 1.80, p\u0026thinsp;=\u0026thinsp;0.171) and vancomycin treatment (OR 2.30, p\u0026thinsp;=\u0026thinsp;0.091) were not significantly correlated with the incidence of kidney deterioration (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eOdds ratio for kidney deterioration\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOdds Ratio (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMale\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.76 (0.4\u0026ndash;1.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.408\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.02 (1.0-1.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.031\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSystolic blood pressure\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.01 (0.99\u0026ndash;1.02)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.418\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSystolic blood pressure\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.97 (0.95-1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSpO\u003c/b\u003e\u003csub\u003e\u003cb\u003e2\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.08 (0.98\u0026ndash;1.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.104\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHeart rate\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.02 (1.01\u0026ndash;1.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.011\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial Creatinine\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.99 (4.89\u0026ndash;16.51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBase creatinine\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.06 (0.02\u0026ndash;0.13)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial bicarbonate\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.96 (0.9\u0026ndash;1.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.257\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial WBC count\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.05 (1.01\u0026ndash;1.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial hemoglobin\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.16 (0.59\u0026ndash;2.29)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.663\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial platelet\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.0 (1.0\u0026ndash;1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.098\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial hematocrit\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.88 (0.7\u0026ndash;1.11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.296\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial sodium level\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.04 (0.98\u0026ndash;1.11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial Potassium level\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.06 (0.67\u0026ndash;1.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.793\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNorepinephrine dose\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.87 (1.3-18.29)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOpen heart surgery\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.69 (0.76\u0026ndash;3.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.196\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCandida endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.0 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBacterial endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e100104412.13 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEndocarditis NOS\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e275887203.26 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRheumatic endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e268265741.5 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eUse of Vancomycin\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.3 (0.87\u0026ndash;6.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.091\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePresence of bacteremia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.8 (0.78\u0026ndash;4.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMSSA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.67 (0.93\u0026ndash;14.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.064\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMRSA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.29 (0.35\u0026ndash;4.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePseudomonas\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.0 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCandidemia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.0 (0.0-inf)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.999\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIntubation at admission\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.76 (0.15\u0026ndash;3.88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.737\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eWBC\u0026thinsp;=\u0026thinsp;white blood cell; MSSA\u0026thinsp;=\u0026thinsp;Methicillin Sensitive Staphylococcus Aureus; MRSA\u0026thinsp;=\u0026thinsp;MSSA\u0026thinsp;=\u0026thinsp;Methicillin Resistant Staphylococcus Aureus; NOS\u0026thinsp;=\u0026thinsp;not otherwise specified\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eMachine Learning Models predicting Kidney deterioration\u003c/h2\u003e \u003cp\u003eIn a 10-fold cross-validation on the training data, LGBM and XGB models performed best, and hyperparameter tuning was conducted, followed by the development of an ensemble model of these two. LGBM, XGB, and the ensemble model showed AUROCs of 0.790, 0.772, and 0.785, respectively, on the test data, all achieving an accuracy of 0.828 (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The DCA plot indicated slightly better performance of XGB over LGBM in the test data (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Initial systolic blood pressure was the second most important variable in LGBM and seventh in XGB, whereas initial diastolic blood pressure was the eighth and fourth most important, respectively. SHAP value plots for both models' predictions on the test data showed a tendency of higher systolic and diastolic blood pressure predicting a lower likelihood of kidney deterioration. Other significant predictors included higher initial creatinine, lower hematocrit, higher initial heart rate, and older age.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance of models predicting kidney function deterioration\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTrain AUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTrain Accuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTrain F1-score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eTrain Recall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTest AUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eTest Accuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eTest F1-score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eTest Recall\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEnsemble\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.922\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.759\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLightGBM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.944\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.819\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.741\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.790\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.194\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eXGBoost\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.925\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.833\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.776\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.772\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.828\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.242\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.148\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eLightGBM\u0026thinsp;=\u0026thinsp;Light Gradient Boosting Machine; XGBoost\u0026thinsp;=\u0026thinsp;Extreme Gradient Boosting; AUC\u0026thinsp;=\u0026thinsp;Area under the curve\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e\u003cbr\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eCausal Inference Deep Learning Model\u003c/h2\u003e \u003cp\u003eThe GANITE model was trained to examine the Average Treatment Effect (ATE) of maintaining mean blood pressure (MBP)\u0026thinsp;\u0026ge;\u0026thinsp;65 mmHg, showing a decrease in the probability of kidney deterioration by 12.7% (95% Confidence Interval [CI]: -14.1 to -11.4%) in the training data, 13.4% (95% CI: -15.4 to -11.4%) in the test data, and 12.9% (95% CI: -14.0 to -11.8%) in the total data (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The GANITE model's AUROC was 0.716 in the test data, with an accuracy of 0.814. Patients undergoing open heart surgery showed a CATE of -15.3% for kidney deterioration at MBP\u0026thinsp;\u0026ge;\u0026thinsp;65mmHg, significantly larger than the \u0026minus;\u0026thinsp;12.1% CATE in patients not undergoing surgery (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Similarly, the CATE for patients over 60 years old was \u0026minus;\u0026thinsp;14.7%, significantly larger than the \u0026minus;\u0026thinsp;11.1% for those 60 years old or younger. Additionally, the effect of MBP\u0026thinsp;\u0026ge;\u0026thinsp;65 mmHg was significantly more pronounced in reducing the risk of kidney deterioration in patients with initial creatinine\u0026thinsp;\u0026gt;\u0026thinsp;2.0 mg/dL and initial heart rate\u0026thinsp;\u0026le;\u0026thinsp;100 beats per minutes. Groups with decreased vs. increased risk of kidney deterioration at MBP\u0026thinsp;\u0026ge;\u0026thinsp;65 showed significant differences in initial creatinine, baseline creatinine, and heart rate (Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAverage treatment effect (ATE) and evaluation indexes of Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE) model\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTrain\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTest\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAccuracy\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.838\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.814\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.831\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAUROC\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.716\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.737\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eATE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.127 (-0.141\u0026ndash;0.114)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.134 (-0.154\u0026ndash;0.114)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.129 (-0.14\u0026ndash;0.118)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eATE\u0026thinsp;=\u0026thinsp;Average treatment effect; AUROC\u0026thinsp;=\u0026thinsp;area under the receiver operating characteristic;\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003econditional average treatment effect (CATE) for total population\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026minus;\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026minus;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCATE for variable positive\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCATE for variable negative\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOpen heart surgery\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.153 (-0.17\u0026ndash;0.135)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.121 (-0.135\u0026ndash;0.108)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.005\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eIntubation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.145 (-0.2\u0026ndash;0.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.128 (-0.14\u0026ndash;0.117)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.541\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBacteremia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.129 (-0.153\u0026ndash;0.106)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.129 (-0.141\u0026ndash;0.117)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.977\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge\u0026thinsp;\u0026gt;\u0026thinsp;60\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.147 (-0.162\u0026ndash;0.133)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.111 (-0.127\u0026ndash;0.094)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eInitial Cr\u0026thinsp;\u0026gt;\u0026thinsp;2mg/dL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.147 (-0.167\u0026ndash;0.127)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.122 (-0.135\u0026ndash;0.109)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHeart Rate\u0026thinsp;\u0026gt;\u0026thinsp;100\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c2\"\u003e \u003cp\u003e-0.068 (-0.089\u0026ndash;0.048)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c3\"\u003e \u003cp\u003e-0.162 (-0.174\u0026ndash;0.151)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eCATE\u0026thinsp;=\u0026thinsp;conditional average treatment effect;Cr\u0026thinsp;=\u0026thinsp;creatinine\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparing negative effect group and positive effect group in total population\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eATE negative\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eATE positive\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOpen Heart Surgery\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.264\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.157\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMale\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.663\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.496\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBacterial Endocarditis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.902\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e59.208 (57.633\u0026ndash;60.782)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e59.69 (54.814\u0026ndash;64.566)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.851\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCreatinine\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.24 (2.004\u0026ndash;2.476)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.633 (1.304\u0026ndash;1.963)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHeart rate\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e92.943 (91.093\u0026ndash;94.793)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e102.578 (95.826-109.329)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBase Creatinine\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.951 (1.74\u0026ndash;2.162)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.36 (1.065\u0026ndash;1.655)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHemoglobin\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e9.962 (9.788\u0026ndash;10.136)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10.042 (9.37-10.714)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.817\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSerum Na\u003c/b\u003e\u003csup\u003e\u003cb\u003e+\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e136.392 (135.889-136.895)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e136.067 (134.398-137.735)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.709\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eATE\u0026thinsp;=\u0026thinsp;Average treatment effect\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis research marks a significant advancement in predicting kidney function deterioration among ICU patients with endocarditis, employing both machine learning models and statistical techniques to understand the factors influencing kidney health. Notably, the study leverages a deep learning causal inference model to demonstrate the protective effect of maintaining a MBP\u0026thinsp;\u0026ge;\u0026thinsp;65 mmHg, underscoring its critical role in mitigating kidney deterioration risk under similar patient conditions.\u003c/p\u003e \u003cp\u003eThe study's findings, particularly the AUROC scores from the test dataset, highlight the predictive power of the LGBM model, underscoring the utility of machine learning in clinical settings. The use of SHAP values from tree-based machine learning models provides deeper insights into how various factors impact kidney deterioration. This approach offers a nuanced understanding beyond traditional logistic regression, which, while effective in illustrating associations between clinical variables and outcomes, may have limitations in assessing model fit and interpretability.\u003c/p\u003e \u003cp\u003eThe differential impact of SBP and DBP on kidney function, as revealed through logistic regression and machine learning analyses, emphasizes the complexity of blood pressure's role in kidney health. The study's progression to analyzing MBP through causal inference analysis addresses this complexity, providing a more holistic view of blood pressure's effects.\u003c/p\u003e \u003cp\u003eThe employment of a deep learning causal inference model represents a methodological strength, offering a way to approximate causality in scenarios where randomization is impractical. This aspect is particularly relevant in the ICU setting, where patient severity and ethical considerations limit the feasibility of randomized controlled trials. The study's ability to infer causality in such a constrained environment adds significant value to its findings, offering a model for future research in similar clinical contexts.\u003c/p\u003e \u003cp\u003eRecent studies underscore the significant risk AKI poses to mortality in endocarditis patients. AKI is a frequent complication during IE, involving 69% of the patients.\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e Known risk factors of AKI is acute heart failure, usage of vancomycin and prosthetic valve involvement. \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e The specific role of antibiotics and their potential renal toxicity in the treatment of infective endocarditis have been documented, indicating the complexity of managing renal health in these patients.\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eAKI is associated with high mortality in IE (22.7% vs. 16.0%).\u003csup\u003e1\u003c/sup\u003e A nationwide study by Petersen et al.\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003erevealed that dialysis-requiring AKI during admission for infective endocarditis is associated with a significantly increased one-year mortality risk from discharge. This finding highlights the critical nature of AKI as a high-risk factor for worse long-term outcomes compared to patients without dialysis requirements.\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e Ortiz-Soriano et al.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e further support this by documenting the high morbidity and mortality associated with AKI in infective endocarditis patients, noting that two-thirds of such patients experience incident AKI, with those suffering from severe AKI incurring increased healthcare costs and a higher risk of mortality.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e These studies collectively emphasize the detrimental impact of AKI on the survival of endocarditis patients, highlighting the urgent need for early detection and intervention to mitigate its effects.\u003c/p\u003e \u003cp\u003eThe significance of maintaining optimal blood pressure for kidney function, especially in septic conditions prevalent among endocarditis patients in the ICU, is well documented. A study by Jamme et al.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e on amoxicillin crystalluria associated with AKI in patients treated for acute infective endocarditis illustrates the complex interplay between treatment modalities for endocarditis and kidney health. \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e This research emphasizes the importance of careful management of medications and supportive treatments to prevent further renal impairment in already at-risk populations. Septic shock and heart failure is well known risk factors of AKI in patients admitted for IE.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThe decision to adopt a mean arterial pressure (MAP) target of \u0026ge;\u0026thinsp;65 mmHg in the management of septic shock, as employed in our study, is firmly grounded in evidence-based guidelines and reinforced by the comprehensive literature review conducted by Leone et al.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. This analysis of seven comparative studies, encompassing both randomized clinical trials and observational studies, supports the adequacy of a 65 mmHg MAP target for most septic shock patients. This threshold is based on the premise that it generally suffices to sustain organ perfusion and function, thereby stabilizing the patient's hemodynamic status.\u003c/p\u003e \u003cp\u003eThe goal of the present study is not to presuppose conclusions but to rigorously investigate the potential protective role of mean blood pressure against kidney function deterioration in ICU-admitted endocarditis patients. Through this investigation, we aim to contribute valuable insights to the existing body of knowledge, facilitating the development of targeted interventions to improve patient outcomes. This approach emphasizes the critical need for comprehensive research into the interplay between renal health, treatment strategies, and patient management in endocarditis, aiming to enhance care and prognosis for this vulnerable patient population\u003c/p\u003e \u003cp\u003eHowever, the study faces limitations, including its relatively small sample size of 484 subjects. While the specificity of the ICU setting and the condition of endocarditis inherently limit the availability of large datasets, this constraint may affect the generalizability of the findings. The absence of certain variables, such as vegetation size and ejection fraction, due to the limitations of the MIMIC-III database, further restricts the study's scope. These omissions highlight the need for more comprehensive datasets in future research to encompass a wider range of clinically relevant variables.\u003c/p\u003e \u003cp\u003eMoreover, while the deep learning causal inference model provides valuable insights, it cannot fully substitute for randomization. This limitation underscores the inherent challenges in drawing causal inferences from observational data, especially in complex clinical settings like the ICU. Future studies might explore innovative methodologies to overcome these challenges, potentially incorporating hybrid models that blend machine learning with traditional epidemiological approaches to better approximate causal relationships.\u003c/p\u003e \u003cp\u003eIn conclusion, this study makes a compelling case for the importance of maintaining MBP\u0026thinsp;\u0026ge;\u0026thinsp;65 in ICU patients with endocarditis to reduce the risk of kidney function deterioration. Its methodological approach, combining statistical techniques with machine learning and deep learning models, offers a blueprint for future research aimed at unraveling the complex interplay of factors affecting patient outcomes in critical care settings.\u003c/p\u003e \u003cp\u003eIn conclusion, the machine learning model was effective in prediction of kidney deterioration in patients with infective endocarditis in ICU. Among these patients, the maintenance of MBP\u0026thinsp;\u0026ge;\u0026thinsp;65mmHg prevented the future kidney function deterioration after ICU admission. This study provides essential insights into kidney health management in critically ill patients, highlighting the potential of advanced analytics in enhancing patient care and outcomes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgement:\u0026nbsp;\u003c/strong\u003eNone\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure:\u003c/strong\u003e No disclosure\u003c/p\u003e\n\u003cp\u003eConflict of interest: None\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding Declaration: None\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declaration: Not applicable\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Declaration: not applicable (publicly available database)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e \u003cstrong\u003eConflict of interest:\u003c/strong\u003e \u003cp\u003eNone\u003c/p\u003e \u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMWK and YK wrote the manuscript and prepared all figures and tables. All authors reviewed the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBor DH, Woolhandler S, Nardin R, Brusch J, Himmelstein DU. Infective endocarditis in the U.S., 1998\u0026ndash;2009: a nationwide study. PLoS One. 2013;8(3):e60033.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConlon PJ, Jefferies F, Krigman HR, Corey GR, Sexton DJ, Abramson MA. Predictors of prognosis and risk of acute renal failure in bacterial endocarditis. Clin Nephrol. 1998;49(2):96\u0026ndash;101.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLegrand M, Pirracchio R, Rosa A, Petersen ML, Van der Laan M, Fabiani JN, et al. Incidence, risk factors and prediction of post-operative acute kidney injury following cardiac surgery for active infective endocarditis: an observational study. Crit Care. 2013;17(5):R220.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson AE, Pollard TJ, Shen L, Lehman LW, Feng M, Ghassemi M, et al. MIMIC-III, a freely accessible critical care database. Sci Data. 2016;3:160035.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePericas JM, Hernandez-Meneses M, Munoz P, Martinez-Selles M, Alvarez-Uria A, de Alarcon A, et al. Characteristics and Outcome of Acute Heart Failure in Infective Endocarditis: Focus on Cardiogenic Shock. Clin Infect Dis. 2021;73(5):765\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGagneux-Brunon A, Pouvaret A, Maillard N, Berthelot P, Lutz MF, Cazorla C, et al. Acute kidney injury in infective endocarditis: A retrospective analysis. Med Mal Infect. 2019;49(7):527\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKunming P, Ying H, Chenqi X, Zhangzhang C, Xiaoqiang D, Xiaoyu L, et al. Vancomycin associated acute kidney injury in patients with infectious endocarditis: a large retrospective cohort study. Front Pharmacol. 2023;14:1260802.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarberan J, Mensa J, Artero A, Epelde F, Rodriguez JC, Ruiz-Morales J, et al. Factors associated with development of nephrotoxicity in patients treated with vancomycin versus daptomycin for severe Gram-positive infections: A practice-based study. Rev Esp Quimioter. 2019;32(1):22\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetersen JK, Jensen AD, Bruun NE, Kamper AL, Butt JH, Havers-Borgersen E, et al. Outcome of Dialysis-Requiring Acute Kidney Injury in Patients With Infective Endocarditis: A Nationwide Study. Clin Infect Dis. 2021;72(9):e232-e9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrtiz-Soriano V, Donaldson K, Du G, Li Y, Lambert J, Cleland D, et al. Incidence and Cost of Acute Kidney Injury in Hospitalized Patients with Infective Endocarditis. J Clin Med. 2019;8(7).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJamme M, Oliver L, Ternacle J, Lepeule R, Moussafeur A, Haymann JP, et al. Amoxicillin crystalluria is associated with acute kidney injury in patients treated for acute infective endocarditis. Nephrol Dial Transplant. 2021;36(10):1955\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHabib G, Lancellotti P, Antunes MJ, Bongiorni MG, Casalta JP, Del Zotti F, et al. 2015 ESC Guidelines for the management of infective endocarditis: The Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J. 2015;36(44):3075\u0026ndash;128.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeone M, Asfar P, Radermacher P, Vincent JL, Martin C. Optimizing mean arterial pressure in septic shock: a critical reappraisal of the literature. Crit Care. 2015;19(1):101.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"infection","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"infe","sideBox":"Learn more about [Infection](http://link.springer.com/journal/15010)","snPcode":"15010","submissionUrl":"https://submission.nature.com/new-submission/15010/3","title":"Infection","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4385746/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4385746/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: Acute kidney injury in infective endocarditis presents significant management challenges in intensive care unit (ICU). We explored the role of mean blood pressure(MBP) at the time of ICU admission predicting kidney function outcomes in endocarditis patients using deep learning model, Generative Adversarial Nets for inference of Individualized Treatment Effects (GANITE).\u003c/p\u003e\n\u003cp\u003eMethods: This study utilized data from the Medical Information Mart for Intensive Care III database. Patients with infective endocarditis admitted to intensive care unit were included in this study. A machine learning model was developed to predict the kidney function deterioration. SHapley Additive exPlanations (SHAP) were used to understand how variables affect kidney function. Moreover, the GANITE model, a causal inference deep learning model, was used to determine the effect of blood pressure to kidney function.\u003c/p\u003e\n\u003cp\u003eResults. A total of 484 patients were included in the analysis, among whom 85(17.6%) experienced kidney deterioration. Light gradient boosting machine, extreme gradient boosting, and the ensemble model showed area under the receiver operating characteristics of 0.790, 0.772, and 0.785, respectively, on the test data, all achieving an accuracy of 0.828. SHAP value plots revealed that higher blood pressure predicted a lower likelihood of kidney deterioration. Analysis using the GANITE model revealed that maintaining MBP≥65mmHg resulted in a decrease in the probability of kidney deterioration by 12.9%.\u003c/p\u003e\n\u003cp\u003eConclusions: In patients with infective endocarditis in ICU, the maintenance of MBP≥65mmHg prevented the future kidney function deterioration after ICU admission.\u003c/p\u003e","manuscriptTitle":"Machine learning based prediction of kidney function deterioration in infective endocarditis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-17 16:31:33","doi":"10.21203/rs.3.rs-4385746/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorAssigned","content":"","date":"2024-05-08T04:43:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-05-08T01:30:15+00:00","index":"","fulltext":""},{"type":"submitted","content":"Infection","date":"2024-05-08T01:25:34+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"infection","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"infe","sideBox":"Learn more about [Infection](http://link.springer.com/journal/15010)","snPcode":"15010","submissionUrl":"https://submission.nature.com/new-submission/15010/3","title":"Infection","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"b09848c3-a379-4935-a192-c3aae84454f1","owner":[],"postedDate":"May 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2024-05-17T16:31:33+00:00","versionOfRecord":[],"versionCreatedAt":"2024-05-17 16:31:33","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4385746","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4385746","identity":"rs-4385746","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}