Comparative Analysis of Machine Learning Models for Prediction of Acute Liver Injury in Sepsis Patients.

Introduction: Acute liver injury (ALI) is a common complication of sepsis and is associated with adverse clinical outcomes. We aimed to develop a model to predict the risk of ALI in patients with sepsis after hospitalization. Methods: Medical records of 3196 septic patients treated at the Lishui Central Hospital in Zhejiang Province from January 2015 to May 2023 were selected. Cohort 1 was divided into ALI and non-ALI groups for model training and internal validation. The initial laboratory test results of the study subjects were used as features for machine learning (ML), and models built using nine different ML algorithms were compared to select the best algorithm and model. The predictive performance of model stacking methods was then explored. The best model was externally validated in Cohort 2. Results: In Cohort 1, LightGBM demonstrated good stability and predictive performance with an area under the curve (AUC) of 0.841. The top five most important variables in the model were diabetes, congestive heart failure, prothrombin time, heart rate, and platelet count. The LightGBM model showed stable and good ALI risk prediction ability in the external validation of Cohort 2 with an AUC of 0.815. Furthermore, an online prediction website was developed to assist healthcare professionals in applying this model more effectively. Conclusions: The Light GBM model can predict the risk of ALI in patients with sepsis after hospitalization.

Application of interpretable machine learning algorithms to predict acute kidney injury in patients with cerebral infarction in ICU.

BACKGROUND: Acute kidney injury (AKI) is not only a complication but also a serious threat to patients with cerebral infarction (CI). This study aimed to explore the application of interpretable machine learning algorithms in predicting AKI in patients with cerebral infarction. METHODS: The study included 3920 patients with CI admitted to the Intensive Care Unit and Emergency Medicine of the Central Hospital of Lishui City, Zhejiang Province. Nine machine learning techniques, including XGBoost, logistics, LightGBM, random forest (RF), AdaBoost, GaussianNB (GNB), Multi-Layer Perceptron (MLP), support vector machine (SVM), and k-nearest neighbors (KNN) classification, were used to develop a predictive model for AKI in these patients. SHapley Additive exPlanations (SHAP) analysis provided visual explanations for each patient. Finally, model effectiveness was assessed using metrics such as average precision (AP), sensitivity, specificity, accuracy, F1 score, precision-recall (PR) curve, calibration plot, and decision curve analysis (DCA). RESULTS: The XGBoost model performed better in the internal validation set and the external validation set, with an AUC of 0.940 and 0.887, respectively. The five most important variables in the model were, in order, glomerular filtration rate, low-density lipoprotein, total cholesterol, hemiplegia and serum kalium. CONCLUSION: This study demonstrates the potential of interpretable machine learning algorithms in predicting CI patients with AKI.

Sodium octanoate alleviates cardiac and cerebral injury after traumatic cardiac arrest in a porcine model.

INTRODUCTION: Traumatic cardiac arrest (TCA) is a severe condition with a high mortality rate, and patients who survive from TCA face a poor prognosis due to post-resuscitation injury, including cardiac and cerebral injury, which remains a serious challenge. Sodium octanoate has shown protective effects against various diseases. The present study aims to investigate sodium octanoate's protective effects against cardiac and cerebral injury after TCA in a porcine model. METHODS: The study included a total of 22 male domestic pigs divided into three groups: Sham group (n = 7), TCA group (n = 7), and sodium octanoate (SO) group (n = 8). Hemorrhage was initiated via the right femoral artery by a blood pump at a rate of 2 ml·kg-1·min-1 to establish TCA model. The Sham group underwent only endotracheal intubation and arteriovenous catheterization, without experiencing the blood loss/cardiac arrest/resuscitation model. At 5 min after resuscitation, the SO group received a continuous sodium octanoate infusion while the TCA group received the same volume of saline. General indicators were monitored, and blood samples were collected at baseline and at different time points after resuscitation. At 24 h after resuscitation, pigs were sacrificed, and heart and brain were obtained for cell apoptosis detection, iron deposition staining, oxidative stress detection, and the expression of ferroptosis-related proteins (ACSL4 and GPX4). RESULTS: Sodium octanoate significantly improved mean arterial pressure, cardiac output and ejection fraction induced by TCA. Serum biomarkers of cardiac and cerebral injury were found to increase at all time points after resuscitation, while sodium octanoate significantly reduced their levels. The apoptosis rates of cardiomyocytes and cerebral cortex cells in the SO group were significantly lower than in the TCA group, along with a reduced area of iron deposition staining. The sodium octanoate also reduced oxidative stress and down-regulated ferroptosis which was indicated by protein level alteration of ACSL4 and GPX4. CONCLUSION: Our study's findings suggest that early infusion of sodium octanoate significantly alleviates post-resuscitation cardiac and cerebral injury in a porcine model of TCA, possibly through inhibition of cell apoptosis and GPX4-mediated ferroptosis. Therefore, sodium octanoate could be a potential therapeutic strategy for patients with TCA.

SULFORAPHANE ALLEVIATES POSTRESUSCITATION LUNG PYROPTOSIS POSSIBLY VIA ACTIVATING THE NRF2/HO-1 PATHWAY.

ABSTRACT: Introduction: Sulforaphane (SFN), known as the activator of the nuclear factor E2-related factor 2 (Nrf2)/heme oxygenase 1 (HO-1) pathway, has been proven to protect the lung against various pathological stimuli. The present study aimed to investigate the effect of SFN on lung injury induced by systemic ischemia reperfusion after cardiac arrest and resuscitation. Methods: After animal preparation, 24 pigs were randomly divided into sham group (n = 6), cardiopulmonary resuscitation group (CPR, n = 9), or CPR + SFN group (n = 9). The experimental model was then established by 10 min of cardiac arrest followed by 6 min of CPR. Once spontaneous circulation was achieved, a dose of 2 mg/kg of SFN diluted in 20 mL of saline was intravenously infused with a duration of 5 min. During 4 h of observation after resuscitation, extravascular lung water index (ELWI), pulmonary vascular permeability index (PVPI), and oxygenation index were regularly evaluated. At 24 h after resuscitation, lung tissues were harvested to evaluate the score of lung histopathological injury, the activity of superoxide dismutase, the contents of malondialdehyde, IL-1ß, and IL-18, and the expression levels of NOD-like receptor pyrin domain 3, cleaved caspase 1, gasdermin D (GSDMD), GSDMD N-terminal, Nrf2, and HO-1. Results: During CPR, spontaneous circulation was achieved in six and seven pigs in the CPR and CPR + SFN groups, respectively. After resuscitation, the indicators of lung injury (ELWI, PVPI, and oxygenation index) were all better in the CPR + SFN group than in the CPR group, in which the differences in ELWI and PVPI at 2, and 4 h after resuscitation were significant between the two groups. In addition, SFN significantly reduced lung injury score, improved oxidative imbalance (superoxide dismutase, malondialdehyde), decreased pyroptosis-related proinflammatory cytokines (IL-1ß, IL-18), downregulated pyroptosis-related proteins (NOD-like receptor pyrin domain 3, cleaved caspase 1, GSDMD, GSDMD N-terminal), and activated the Nrf2/HO-1 pathway when compared with the CPR group. Conclusion: SFN produced effective postresuscitation lung protection through alleviating lung pyroptosis possibly via activating the Nrf2/HO-1 pathway in pigs.

ALDA-1 TREATMENT ALLEVIATES LUNG INJURY AFTER CARDIAC ARREST AND RESUSCITATION IN SWINE.

ABSTRACT: Introduction: Alda-1, an aldehyde dehydrogenase 2 (ALDH2) activator, has been shown to protect the lung against a variety of diseases including regional ischemia-reperfusion injury, severe hemorrhagic shock, hyperoxia, and so on. The present study was designed to investigate the effectiveness of Alda-1 treatment in alleviating lung injury after cardiac arrest (CA) and cardiopulmonary resuscitation (CPR) in swine. Methods: A total of 24 swine were randomized into three groups: sham (n = 6), CA/CPR (n = 10), and CA/CPR + Alda-1 (n = 8). The swine model was established by 8 min of electrically induced and untreated CA, and then 8 min of manual CPR. A dose of 0.88 mg/kg of Alda-1 was intravenously injected at 5 min after CA/CPR. After CA/CPR, extravascular lung water index (ELWI), pulmonary vascular permeability index (PVPI), and oxygenation index (OI) were regularly evaluated for 4 h. At 24 h after resuscitation, lung ALDH2 activity was detected, and its injury score, apoptosis, and ferroptosis were measured. Results: After experiencing the same procedure of CA and CPR, five swine in the CA/CPR group and six swine in the CA/CPR + Alda-1 group restored spontaneous circulation. Subsequently, significantly increased ELWI and PVPI, and markedly decreased OI were observed in these two groups compared with the sham group. However, all of them were gradually improved and significantly better in the swine treated with the Alda-1 compared with the CA/CPR group. Tissue analysis indicated that lung ALDH2 activity was significantly decreased in those swine experiencing the CA/CPR procedure compared with the sham group; nevertheless, its activity was significantly greater in the CA/CPR + Alda-1 group than in the CA/CPR group. In addition, lung injury score, and its apoptosis and ferroptosis were significantly increased in the CA/CPR and CA/CPR + Alda-1 groups compared with the sham group. Likewise, Alda-1 treatment significantly decreased these pathological damages in lung tissue when compared with the CA/CPR group. Conclusions: Alda-1 treatment was effective to alleviate lung injury after CA/CPR in a swine model, in which the protective role was possibly related to the inhibition of cell apoptosis and ferroptosis. It might provide a novel therapeutic target and a feasible therapeutic drug for lung protection after CA/CPR.