Daily Ards Research Analysis
Analyzed 6 papers and selected 3 impactful papers.
Summary
Three studies advance ARDS-related care across the perinatal and critical care spectrum: a multicenter cluster-randomized crossover trial protocol aims to standardize initial oxygen for moderate–late preterm infants; a MIMIC-IV machine learning model predicts 28-day mortality in hepatic failure complicated by ARDS; and a small cohort suggests VV-to-venopulmonary ECMO conversion may stabilize hemodynamics and end-organ function in refractory respiratory failure.
Research Themes
- Risk stratification and prognostication in ARDS
- Optimization of initial respiratory support in preterm infants
- Advanced extracorporeal support strategies for refractory respiratory failure
Selected Articles
1. Initial supplementary oxygen concentration for moderate-late preterm infants receiving respiratory support in the delivery room: study protocol for the multicenter, cluster-randomized, crossover AIROPLANE trial.
Protocol for a pragmatic, multicenter cluster-randomized crossover RCT comparing 30% vs 21% initial oxygen for moderate–late preterm infants needing delivery-room respiratory support. The trial (≥1200 infants, ≥20 Australian sites) targets the primary outcome of ongoing respiratory support on leaving the delivery room and aims to inform future international guidelines.
Impact: First randomized study focused exclusively on moderate–late preterm infants to address an evidence gap in delivery-room oxygen strategy with a robust, pragmatic design.
Clinical Implications: If 30% or 21% oxygen proves superior, delivery-room protocols for moderate–late preterm infants could be standardized, potentially reducing oxygen-related harm or treatment failure.
Key Findings
- Pragmatic multicenter cluster-randomized crossover RCT comparing initial 30% vs 21% oxygen for 32–35+6 week infants.
- Primary outcome: need for ongoing respiratory support upon leaving the delivery room.
- Planned enrollment ≥1200 infants across ≥20 Australian sites; trial registered (ACTRN12621001267842).
Methodological Strengths
- Cluster-randomized crossover design reduces contamination and reflects real-world practice.
- Large planned sample size with multicenter participation and pre-registration.
Limitations
- Protocol paper: no outcomes reported yet.
- Unblinded design may introduce performance bias; site heterogeneity likely.
Future Directions: Complete enrollment and report clinical outcomes, including safety (hypoxemia/hyperoxia), and perform subgroup analyses (GA strata, mode of support).
BACKGROUND: Moderate-late preterm infants born at 32-35 completed weeks' gestation constitute a large proportion of all preterm births (< 37-week gestation), yet they are not well represented in the newborn resuscitation literature. Preterm infants often receive respiratory support in the delivery room, and recommendations exist to guide the use of supplemental oxygen when providing this support. However, there are minimal data regarding the best initial supplementary oxygen concentration for moderate-late preterm infants requiring respiratory support at birth, resulting in practice variation. The aim of this trial is to compare initial supplementary oxygen concentrations of 30% and 21% (air) in preterm infants of 32-35 weeks' gestation who require respiratory support in the delivery room, with a primary outcome of the need for ongoing respiratory support upon leaving the delivery room. METHODS: This is a prospective, unblinded, multicenter, cluster-randomized, crossover trial in Australian maternity hospitals comparing initial supplementary oxygen concentrations of 30% and 21% (air) in moderate-late preterm infants who require respiratory support at birth. Eligible infants are those born from 32 + 0 to 35 + 6 weeks' gestation without major cardiorespiratory or craniofacial anomalies, who are receiving active care, and who receive respiratory support in the delivery room within the first three minutes after birth. The primary outcome is the need for ongoing respiratory support upon leaving the delivery room. The trial will recruit a minimum of 1200 infants from at least 20 study sites in Australia using a waiver of consent process. DISCUSSION: The AIROPLANE trial is a pragmatic study in an underrepresented population using a novel study design. It will be the first randomized study of initial supplementary oxygen concentration during respiratory support in the delivery room to exclusively recruit moderate-late preterm infants. The results will address an important evidence gap and will inform future international guidelines. TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry ACTRN12621001267842. Registered on 20th September 2021.
2. Machine learning prediction model for 28-day mortality among hepatic failure patients complicated by acute respiratory distress syndrome.
Using MIMIC-IV data, a random forest model predicted 28-day mortality in hepatic failure patients with ARDS (AUC 0.823). SHAP highlighted eight clinical predictors, potentially enabling earlier risk stratification and targeted management.
Impact: Provides an interpretable ML tool for a high-risk ARDS subgroup (hepatic failure), with strong internal validation and clinically intuitive predictors.
Clinical Implications: Can support triage and early escalation (e.g., monitoring, ventilatory strategies, renal/hemodynamic support) for high-risk patients; may inform resource allocation in ICUs.
Key Findings
- Included 884 hepatic failure patients with concurrent ARDS; 28-day mortality was 47.4%.
- Random forest achieved AUC 0.823 (95% CI 0.763–0.883) in the validation set, outperforming other ML methods.
- SHAP identified key predictors: age, neutrophil count, pulse transit time, direct bilirubin, heart rate, fibrinogen, serum sodium, and prothrombin time.
Methodological Strengths
- Large ICU dataset with predefined train/validation split and comparison of multiple ML algorithms.
- Model interpretability via SHAP enhances clinical relevance and transparency.
Limitations
- Retrospective single-database study with internal validation only; no external or prospective validation.
- Potential residual confounding, missing data, and center-specific practice patterns in MIMIC-IV.
Future Directions: External validation across diverse ICUs, prospective impact studies, and integration into clinical workflows with decision-support thresholds.
OBJECTIVES: Hepatic failure is a common and severe condition among intensive care unit (ICU) patients. Its complication with acute respiratory distress syndrome (ARDS) is consistently associated with poor clinical outcomes and a significant disease burden. Early identification of high-risk patients is essential for improving clinical outcomes. This study aimed to develop and validate a machine learning (ML) model to predict 28-day mortality in ICU patients with hepatic failure complicated by ARDS. METHODS: Data were extracted from the Medical Information Mart for Intensive Care IV database, focusing on patients with hepatic failure complicated by ARDS. The cohort was randomly divided into an 80% training set and a 20% validation set. Six ML algorithms were applied to analyze clinical characteristics. Shapley Additive Explanations (SHAP) were used to interpret the optimal model. RESULTS: A total of 884 patients with hepatic failure and concurrent ARDS were included, with a 28-day mortality rate of 47.4%. Random forest models demonstrated superior performance, achieving an area under the curve of 0.823 (95% confidence interval: 0763-0.883) in the validation set. SHAP analysis identified eight clinically significant predictors of mortality, ranked by importance: age, neutrophil count, pulse transit time, direct bilirubin, heart rate, fibrinogen, serum sodium concentration, and prothrombin time. SHAP enhanced model interpretability, supporting clinical decision-making and potentially improving patient outcomes. CONCLUSIONS: ML approaches exhibited promising performance in predicting 28-day mortality among hepatic failure patients complicated by ARDS. These models may aid in guiding treatment decisions for patients with hepatic failure patients.
3. Outcomes of Venopulmonary Extracorporeal Membrane Oxygenation Conversion in Patients With Severe Respiratory Failure on Venovenous Support.
In 19 adults on VV ECMO with refractory RVD, conversion to VP ECMO was associated with improved hemodynamics and oxygenation, AKI resolution in 62.5% with pre-existing AKI, and reduced pressor needs in 78%. Findings support considering VP configuration in selected refractory cases.
Impact: Provides pragmatic evidence on configuration change (VV to VP ECMO) for refractory right heart compromise during severe respiratory failure—a critical management question with limited data.
Clinical Implications: ECMO centers may consider VP conversion in VV ECMO patients with refractory RVD to stabilize hemodynamics and support organ function, alongside careful patient selection and monitoring.
Key Findings
- Retrospective cohort of 19 adults on VV ECMO converted to VP ECMO; outcomes included mortality, AKI, pressor needs, and ECMO/ventilator parameters.
- AKI resolution occurred in 62.5% (5/8) with pre-conversion AKI; 40% (2/5) liberated from CRRT.
- Post-conversion reductions observed in pressor requirements (78%), ECMO flows (79%), sweep gas flow (63%), ECMO FiO2 (26%), and ventilator FiO2 (16%); sedation weaning feasible in 53%.
Methodological Strengths
- Granular physiological and support parameter reporting before and after configuration change.
- Focus on a clinically challenging phenotype (refractory RVD on VV ECMO).
Limitations
- Small single-center retrospective cohort without a control group limits causal inference.
- Mortality impact not clearly quantified in the abstract; potential selection and survivorship biases.
Future Directions: Prospective multicenter registries or trials to define indications, timing, and outcomes of VP conversion versus alternative strategies.
Severe respiratory failure is frequently complicated by right ventricular dysfunction (RVD), which occurs in 20-50% of cases. In patients on venovenous extracorporeal membrane oxygenation (VV ECMO) with refractory RVD, conversion to venopulmonary (VP) ECMO can provide additional mechanical support. This study evaluates the impact of VV to VP ECMO conversion on mortality and end-organ dysfunction in severe respiratory failure. A retrospective cohort study of 19 adult patients on VV ECMO who were converted to VP ECMO was performed. Outcomes included in-hospital mortality, resolution of acute kidney injury (AKI), pressor requirements, ventilator and ECMO parameters, and ECMO support duration. Venopulmonary ECMO conversion facilitated AKI resolution in 62.5% of patients with pre-conversion AKI (5/8) and was associated with liberation from continuous renal replacement therapy in 40% of patients (2/5). Conversion resulted in reduced pressor requirements (7/9, 78%), ECMO flows (15/19, 79%), sweep gas flow (12/19, 63%), ECMO fraction of inspired oxygen (FiO2; 5/19, 26%), and ventilator FiO2 (3/19, 16%). Venopulmonary ECMO conversion facilitated sedation weaning in 10 patients (53%) which was previously not tolerated due to desaturation events on VV ECMO. Overall, VP ECMO conversion was associated with improved oxygenation, hemodynamic stability, and end-organ function in the majority of patients.