Daily Ards Research Analysis
Among ARDS-related studies, a regional analysis of the phase 3 PANAMO RCT reports mortality reduction with the C5a inhibitor vilobelimab in intubated COVID-19 patients in Western Europe, while effects varied elsewhere. A large MIMIC-IV cohort suggests early acetaminophen after CABG-related ARDS is associated with lower early mortality and shorter ventilation. Machine-learning models predicted in-hospital mortality in sepsis-related ARDS, highlighting APACHE III, bicarbonate, anion gap, and systo
Summary
Among ARDS-related studies, a regional analysis of the phase 3 PANAMO RCT reports mortality reduction with the C5a inhibitor vilobelimab in intubated COVID-19 patients in Western Europe, while effects varied elsewhere. A large MIMIC-IV cohort suggests early acetaminophen after CABG-related ARDS is associated with lower early mortality and shorter ventilation. Machine-learning models predicted in-hospital mortality in sepsis-related ARDS, highlighting APACHE III, bicarbonate, anion gap, and systolic NIBP as key features.
Research Themes
- Complement inhibition in severe COVID-19 ARDS
- Repurposing common analgesics in post-CABG ARDS
- Machine-learning risk stratification in sepsis-related ARDS
Selected Articles
1. Regional comparison of efficacy and safety for vilobelimab in critically ill, invasively mechanically ventilated COVID-19 patients.
In a prespecified regional analysis of the phase 3 PANAMO RCT in intubated COVID-19 patients, vilobelimab reduced 28-day mortality in Western Europe (21% vs 37%; HR 0.51) with consistent safety, while effects were not significant in South America and South Africa/Russia. Age imbalance in Brazil may explain regional heterogeneity.
Impact: This study refines the understanding of where complement C5a blockade provides mortality benefit in critically ill, invasively ventilated COVID-19, informing population selection for future trials and clinical use.
Clinical Implications: Consider C5a blockade with vilobelimab for invasively ventilated COVID-19 patients where healthcare contexts mirror Western Europe; interpret regional effects cautiously and assess patient selection and system factors.
Key Findings
- 28-day mortality was lower with vilobelimab in Western Europe (21% vs 37%; HR 0.51; p=0.014).
- No significant mortality difference in South America (40% vs 37%; HR 0.94; p=0.83) or South Africa/Russia (69% vs 87%; HR 0.62; p=0.25).
- Safety profiles were similar across regions; Brazilian subgroup showed age imbalance (older in vilobelimab arm).
Methodological Strengths
- Phase 3, randomized, double-blind, placebo-controlled, multicenter design
- Prespecified regional pooling with time-to-event analyses (HRs) and consistent safety monitoring
Limitations
- Regional analyses, while prespecified, are subgroup evaluations and may be underpowered outside Western Europe
- Age imbalance in the Brazilian subgroup and COVID-19-era practice variability may confound regional comparisons
Future Directions: Validate findings in independent cohorts and explore determinants of regional effect modification; assess C5a blockade in non-COVID ARDS and in enriched phenotypes.
BACKGROUND: Vilobelimab, a first in class C5a-specific monoclonal antibody, improved 28-day and 60-day mortality in intubated COVID-19 patients in PANAMO, a phase 3 randomised, double-blind, placebo-controlled multicentre study. All-cause mortality was pre-specified to be analysed pooling by region (western Europe, South America, South Africa/Russia). METHODS: Critically ill, invasively mechanically ventilated COVID-19 patients were randomised in a 1:1 ratio within 48 hours of intubation to receive vilobelimab treatment (six, 800 mg intravenous infusions) or placebo on top of standard of care. We analysed the efficacy and safety of vilobelimab based on prespecified geographic regions. RESULTS: 368 patients were randomised and analysed: 177 in the vilobelimab group and 191 in the placebo group. In western Europe (n=209), 28-day all-cause mortality was significantly lower in the vilobelimab group (21%) compared with placebo (37%) (HR 0.51 (95% CI: 0.30, 0.87), p=0.014). In South America (n=126), mortality was similar between groups (40% vs 37%; HR 0.94 (95% CI: 0.53, 1.67), p=0.83). In South Africa/Russia (n=33), mortality was 69% in the vilobelimab group and 87% in the placebo group (HR 0.62 (95% CI: 0.28, 1.38), p=0.25). Within the Brazilian subpopulation (n=74), a significant age imbalance between the vilobelimab and placebo group was detected (median 53.5 years in the vilobelimab group vs 44.5 years in the placebo group). Occurrence of treatment-emergent adverse events between regions was similar. CONCLUSION: The most apparent 28-day all-cause mortality benefit for vilobelimab was in western Europe. Age imbalance between treatment groups in Brazil may have resulted in a lower efficacy signal for vilobelimab in South America compared with other regions. Overall, vilobelimab demonstrated a favourable safety profile and reduced mortality in critically ill, intubated COVID-19 patients, with regional variations influencing outcomes.
2. Early acetaminophen administration is associated with lower mortality among ARDS patients after coronary artery bypass grafting: a retrospective study.
In a large retrospective MIMIC-IV cohort of CABG-related ARDS, early acetaminophen exposure correlated with lower 14-day mortality (0.5% vs 2.7%; OR 0.301) and shorter hospital stay and mechanical ventilation, with robustness across IPTW, OW, and PSM, and consistency up to 90 days.
Impact: Identifies a low-cost, widely available therapy associated with improved outcomes in a high-risk ARDS subgroup, supporting rapid prospective evaluation.
Clinical Implications: Consider acetaminophen early after CABG-ARDS as part of multimodal care while awaiting RCTs; monitor for confounding and individual risk factors when interpreting effects.
Key Findings
- Early acetaminophen exposure was associated with lower 14-day mortality (0.5% vs 2.7%; OR 0.301; p<0.001).
- Survival benefits remained in Cox analysis (HR 0.329; p<0.001) and across IPTW, OW, and PSM.
- Shorter hospital length of stay and reduced duration of mechanical ventilation (both p<0.001).
- Consistency in 30-, 60-, and 90-day mortality analyses.
Methodological Strengths
- Large sample size (N=5459) with multivariable adjustment and survival analysis
- Multiple causal inference approaches (IPTW, overlap weighting, propensity score matching) to test robustness
Limitations
- Retrospective design with potential residual confounding and confounding by indication
- Single database (MIMIC-IV) may limit generalizability; dosing/timing details and physiological confounders may be incomplete
Future Directions: Conduct randomized trials to test acetaminophen in CABG-related ARDS and explore mechanisms (e.g., antipyretic/anti-inflammatory effects) and optimal timing/dose.
BACKGROUND: Acetaminophen (APAP) is widely used in the treatment of patients after surgery, but the prognosis of patients with coronary artery bypass grafting (CABG)-related acute respiratory distress syndrome (CABG-ARDS) is still unclear. This study aims to explore the role of APAP in the management of CABG related ARDS. METHODS: We collected clinical data on patients with CABG-ARDS from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. The primary outcome was early mortality after ARDS, and the secondary outcomes were length of hospital stay and duration of mechanical ventilation (MV). Multivariate logistic regression and Cox regression models were used for statistical analysis, and inverse probability processing weighting (IPTW), overlap weighting (OW) and propensity score matching (PSM) were used to explore the robustness of the outcomes. RESULTS: A total of 5459 patients were enrolled in the analysis. Multivariate logistic regression analysis revealed that the 14-day mortality in APAP group was significantly lower than that in non-APAP group (0.5% vs. 2.7%, OR = 0.301; 95% CI, 0.170-0.531; P < 0.001). The APAP group also showed a significant advantage in Cox regression analysis (0.5% vs. 2.7%, HR = 0.329; 95% CI, 0.187-0.577; P < 0.001). IPTW, OW, and PSM analyses were conducted between the two groups, and the differences remained significant. These results were consistent in 30-, 60-, and 90-day mortality analyses. Meanwhile, exposure to APAP was associated with a shorter length of hospital stay and a reduced duration of MV (P < 0.001). CONCLUSION: The administration of APAP was associated with reduced early mortality in patients with CABG-ARDS, as well as shorter length of hospital stay and duration of MV.
3. Predicting mortality and risk factors of sepsis related ARDS using machine learning models.
Using 3,386 sepsis-related ARDS cases from MIMIC, multiple ML models predicted in-hospital mortality; the random forest achieved the best test AUROC (0.846). Feature importance highlighted APACHE III, bicarbonate, anion gap, and systolic NIBP as key risk factors.
Impact: Provides an internally validated ML tool and transparent risk features for sepsis-related ARDS, supporting risk stratification and potential clinical decision support.
Clinical Implications: ML-based risk scores could aid triage and personalized management in sepsis-related ARDS, prioritizing high-risk patients and informing goals-of-care discussions, pending external validation.
Key Findings
- Random forest achieved the best test-set performance (AUROC 0.846; 95% CI 0.818–0.874).
- Top predictors included APACHE III, bicarbonate, anion gap, and systolic non-invasive blood pressure.
- Training-set AUCs of 1.0 for RF and LightGBM indicate strong fit but underscore overfitting risk without external validation.
Methodological Strengths
- Large cohort with clear inclusion/exclusion and 70/30 train-test split
- Model comparison across six algorithms with VIF collinearity checks and partial dependence analyses
Limitations
- No external validation; generalizability beyond MIMIC is uncertain
- Potential overfitting (perfect training AUROC for RF/LightGBM) and reliance on available variables with possible missingness
Future Directions: Perform external, multicenter validation and prospective impact studies; integrate models into EHR workflows with real-time calibration and fairness assessment.
Sepsis related acute respiratory distress syndrome (ARDS) is a common and serious disease in clinic. Accurate prediction of in-hospital mortality of patients is crucial to optimize treatment and improve prognosis under the new global definition of ARDS. Our study aimed to use machine learning models to develop models that can effectively predict the in-hospital mortality of patients with sepsis related ARDS, calculate the mortality, and to identify related risk factors under the new global definition of ARDS. Based on MIMIC database, our study included 3470 first-time admission records of patients with sepsis related ARDS. After excluding 4 patients under the age of 18, 75 patients with less than 24 h stay in ICU, and 5 cases with missing indicators > 30%, finally 3386 cases were retained. The variance inflation factor (VIF) analysis was used to test the collinearity of the explanatory variables. The data were divided into the training set and the test set according to the ratio of 7:3. Six models, extreme gradient boosting (XGBoost), light gradient boosting (LightGBM), random forest (RF), classification and regression tree (CART), naive bayes (NB) and logistic regression (LR), were designed for training and testing. In the training set, XGBoost (AUROC = 0.951, 95% CI 0.942-0.961), LR (AUROC = 0.835, 95% CI 0.817-0.854), RF (AUROC = 1.0, 95% CI 1.0-1.0), LightGBM (AUROC = 1.0, 95% CI 1.0-1.0), CART (AUROC = 0.831, 95% CI 0.811-0.852), NB (AUROC = 0.793, 95% CI 0.772-0.814). In the test set, XGBoost (AUROC = 0.833, 95% CI 0.804-0.861), LR (AUROC = 0.82695% CI 0.796-0.856), RF (AUROC = 0.846, 95% CI 0.818-0.874), LightGBM (AUROC = 0.827, 95% CI 0.798-0.856), CART (AUROC = 0.753, 95% CI 0.718-0.787), NB (AUROC = 0.799, 95% CI 0.768-0.831). The RF model has the best performance on the test set. Further analyze the feature importance ranking and partial dependence plots of random forest model. Acute physiology and chronic health evaluation III (APACHE III), bicarbonate, anion gap and non-invasive blood pressure systolic were identified as the four most important risk characteristics. In this study, a variety of machine learning models have been successfully constructed to predict the in-hospital mortality of patients with sepsis related ARDS, among which the RF model performs well. Key risk factors identified include APACHE III, bicarbonate, anion gap and non-invasive blood pressure systolic. The identification of these factors helps clinicians to assess patients' conditions more accurately and develop personalized treatment plans, thereby improving the survival rate and prognosis quality of patients under the new global definition of ARDS.