Daily Sepsis Research Analysis

BACKGROUND: Septic shock is a severe and life-threatening complication of sepsis associated with high mortality. Early identification remains challenging due to the heterogeneous clinical presentation of patients, incomplete real-world ICU data, and the dynamic pathophysiological progression of sepsis. This study aimed to develop and externally validate machine learning (ML) models to predict the progression to septic shock in intensive care unit (ICU) patients with sepsis. METHODS: Data were extracted from two large critical care databases: MIMIC-IV (training set) and eICU-CRD (validation set). Variable selection was performed using LASSO regression. Six ML algorithms-random forest (RF), XGBoost, support vector machine, light gradient boosting machine, logistic regression, and naïve Bayes-were trained on the MIMIC-IV dataset and externally validated on the eICU dataset. Model performance was evaluated using AUC, F1 score, sensitivity, specificity, and balanced accuracy. Model interpretability was assessed using SHapley Additive exPlanations (SHAP). RESULTS: A total of 11,383 septic patients were included, of whom 2,259 in the training cohort and 1,212 in the validation cohort developed septic shock. The RF model achieved the best performance, with an AUC of 0.785, balanced accuracy of 0.717, and an F1 score of 0.511. SHAP analysis identified SOFA score, heart rate, creatinine, SAPS II, and OASIS as the most influential predictors. CONCLUSION: The proposed ML model enables early prediction of septic shock using routinely collected ICU data. SHAP-based interpretation enhances transparency and clinical interpretability. This approach may assist in timely risk stratification, support data-driven decision-making, and ultimately improve outcomes in patients with sepsis.

BACKGROUND AND AIMS: Critically ill patients are believed to experience a dynamic progression of energy metabolism according to the severity and phase of the illness. Although optimizing nutrition and physical therapy to the metabolic profile is recommended for better outcomes, the mechanisms for disease-related metabolic changes remain unclear. This study aimed to elucidate key metabolites and pathways associated with time-course metabolic changes and clinical parameters in acute-phase critically ill patients using untargeted metabolomics. METHODS: We conducted a single-center prospective case series study on critically ill adults expected to require mechanical ventilation for at least 7 days in our intensive care unit. Data collection was started within 48 h of ICU admission, and daily serum samples from day 1 to day 7 were collected. Untargeted metabolomics was performed using liquid chromatography/mass spectrometry. Statistical analyses included principal component analysis, (orthogonal) partial least squares discriminant analysis, and pathway analysis. RESULTS: Ten patients were analyzed during the study period from July 2021 to September 2022. A total of 123 metabolites were annotated by untargeted metabolomics, with significant time-course changes in galactonic acid, ornithine, and l-arginine. Pathway analysis indicated alterations in the arginine biosynthesis pathway. A subgroup analysis showed distinct metabolic profiles for sepsis and non-sepsis patients, with creatine phosphate, uric acid, and creatinine being significant markers. In sepsis patients, metabolic changes strongly correlated with the sequential organ failure assessment (SOFA) score. CONCLUSION: Using untargeted metabolomics, we annotated several metabolites and metabolic pathways strongly associated with time-course changes in the metabolic profile. In addition, it is suggested that nutritional therapy can be optimized according to specific pathophysiology.

BACKGROUND: Bacteremia is commonly present in patients with septic shock. Previous studies have found mixed associations between hospital outcomes and the presence of bacteremia in this patient population. OBJECTIVES: The primary aim of this study was to determine whether adult patients who present to the emergency department (ED) with suspected septic shock and who have bacteremia from ED-derived blood cultures have differences in outcomes. Secondary aims included the association of Gram-positive, Gram-negative, and mixed Gram-positive and Gram-negative bacteremia on hospital mortality. Finally, we sought to create a model to predict bacteremia based on ED variables. METHODS: We conducted a retrospective observational analysis at an urban academic center over a 5-year period between 2013 and 2018. Suspected septic shock was defined by the following: blood cultures collected, antibiotics administered, and vasopressors initiated in the ED. Hospital outcomes were compared in those with and without bacteremia. A logistic regression analysis was used to fit a model to best predict the presence of bacteremia based on ED variables. RESULTS: We analyzed 470,558 patient encounters, and 847 were classified as suspected septic shock, meeting inclusion criteria. Two hundred forty-three (28.7%) were categorized as having bacteremia. No statistical differences were detected in univariate comparisons between patients with and without bacteremia in hospital mortality, hospital length of stay, intensive care unit length of stay, or intubation during hospitalization between those with and without bacteremia. Hospital mortality was not associated with bacteremia based on subgroup analysis by Gram stain. CONCLUSIONS: Adult patients who present to the ED with suspected septic shock and have initial blood cultures that result in bacteremia experience hospital outcomes comparable with those without bacteremia.