Background: Sepsis-associated acute kidney injury (S-AKI) has a significant impact on patient survival, with continuous renal replacement therapy (CRRT) being a crucial intervention. However, the optimal timing for CRRT initiation remains controversial.Methods: Using the MIMIC-IV database for model development and the eICU database for external validation, we analyzed patients with S-AKI to compare survival rates between early and late CRRT initiation groups. Propensity score matching was performed to address potential selection bias. Subgroup analyses stratified patients by disease severity using SOFA scores (low ≤10, medium 11–15, high >15) and creatinine levels (low ≤3 mg/dL, medium 3–5 mg/dL, high >5 mg/dL). Multiple machine learning models were developed and evaluated to predict patient prognosis, with Shapley Additive exPlanations (SHAP) analysis identifying key prognostic factors.Results: After propensity score matching, late CRRT initiation was associated with improved survival probability, but led to increased hospital and ICU stays. Subgroup analyses showed consistent trends favoring late CRRT across all SOFA categories, with the most pronounced effect in high SOFA scores (>15, p = 0.058). The GBM model demonstrated robust predictive performance (average C-index 0.694 in validation and test sets). SHAP analysis identified maximum lactate levels, age, and minimum SpO2 as the strongest predictors of mortality, while CRRT timing showed relatively lower impact on outcome prediction.Conclusion: While later initiation of CRRT in S-AKI patients was associated with improved survival, this benefit comes with increased healthcare resource utilization. The clinical parameters, rather than CRRT timing, are the primary determinants of patient outcomes, suggesting the need for a more personalized approach to CRRT initiation based on overall illness severity.

Background: Bronchial rupture in children is a rare but dangerous complication after chest trauma and is associated with increased mortality. Veno-venous (V-V) extracorporeal membrane oxygenation (ECMO) is reported as one of the treatments for this life-threatening complication. Case presentation: A 4-year-old boy with bronchial rupture and traumatic wet lung complicated by cardiac arrest after chest trauma was admitted to an adult ECMO centre. He experienced two cardiac arrests, one before and one during the operation. The total duration of cardiac arrest was 30 min. V-V ECMO was initiated because of severe hypoxia and hypercapnia during the operation. ECMO was performed for 6 days, and mechanical ventilation lasted 11 days. On the 31st day after surgery, he had recovered completely and was discharged without neurological deficit. Conclusion: V-V ECMO can be considered for supportive care in children with severe acute respiratory failure after bronchial rupture. In an emergency, V-V ECMO can be carried out effectively in a qualified and experienced adult ECMO centre. However, the application of ECMO in children is different from that in adults and requires more refined management.

Extracorporeal membrane oxygenation (ECMO) is a critical life-support intervention for patients with severe cardiopulmonary failure. However, its use is associated with a substantially increased risk of nosocomial infections, with reported incidence rates ranging from 8.8% to 64.0%. These infections—particularly ventilator-associated pneumonia and bloodstream infections—are linked to heightened morbidity, prolonged intensive care and hospital stays, and elevated mortality. This review aims to systematically compile Chinese and English literature published between 2018 and 2025, clarify the unique pathophysiological mechanisms of ECMO-related infections, analyze the limitations and breakthroughs of existing prediction models, and explore the potential role of machine learning in developing personalized early warning systems. Additionally, it seeks to establish a clinical decision-making framework for precise prevention and control. We conclude that improving ECMO infection control requires establishing standardized, clinically applicable diagnostic criteria, conducting a multicenter prospective validation study, and developing transparent, AI-enhanced predictive tools to enable real-time infection monitoring and improved patient prognosis during ECMO support.

SOFA-1 underpins Sepsis-3 definitions but may no longer optimally reflect contemporary ICU practice; SOFA-2 was recently developed from large multinational datasets, yet its prognostic performance in septic populations remains unknown. We aimed to compare the prognostic performance of SOFA-1 and SOFA-2 in ICU patients with sepsis. Retrospective cohort study. This study was conducted using the MIMIC-IV database. A total of 30,667 adult patients (≥18 years) with first ICU admissions meeting Sepsis-3 criteria were included. ICU mortality and in-hospital mortality. We compared discrimination (AUC), calibration, net reclassification improvement (NRI), integrated discrimination improvement (IDI), decision curve analysis and time-dependent survival. Among 30,667 patients, 87.0% met both SOFA-1 and SOFA-2 criteria, 12.0% met only SOFA-2 criteria, and 1.0% met only SOFA-1 criteria. SOFA-2 total scores were systematically higher than SOFA-1, driven primarily by updated renal, cardiovascular and brain scoring criteria. SOFA-2 demonstrated modestly but consistently better discrimination for ICU mortality (AUC 0.779 vs. 0.755; P < 0.001) and hospital mortality (AUC 0.745 vs. 0.725; P < 0.001), with significant reclassification improvement for ICU mortality (NRI 0.143; IDI 0.020). In multivariable Cox models, SOFA-2 achieved higher AUCs for 7-day (0.790 vs. 0.771), 28-day (0.709 vs. 0.697) and 90-day mortality (0.682 vs. 0.671; all P < 0.001), with consistent advantages across all prespecified subgroups. SOFA-2 provides modestly but consistently better prognostic performance than SOFA-1 in septic ICU patients. Whether this incremental gain justifies the added scoring complexity warrants prospective evaluation.