Daily Respiratory Research Analysis
Three high-impact studies in respiratory medicine stood out today: a network meta-analysis of 25 RCTs clarifies comparative benefits and risks across surgical and bronchoscopic lung volume reduction strategies for severe emphysema; a systematic review/meta-analysis shows metagenomic next-generation sequencing in lower respiratory tract infections increases appropriate antibiotic changes and may reduce in-hospital mortality; and a large multimodal deep-learning model (PreEMPT-ECMO) enables contin
Summary
Three high-impact studies in respiratory medicine stood out today: a network meta-analysis of 25 RCTs clarifies comparative benefits and risks across surgical and bronchoscopic lung volume reduction strategies for severe emphysema; a systematic review/meta-analysis shows metagenomic next-generation sequencing in lower respiratory tract infections increases appropriate antibiotic changes and may reduce in-hospital mortality; and a large multimodal deep-learning model (PreEMPT-ECMO) enables continuous prediction of ECMO utilization up to 96 hours, supporting triage and resource allocation.
Research Themes
- Comparative effectiveness of lung volume reduction strategies in severe emphysema
- Pathogen-agnostic sequencing to guide therapy in lower respiratory tract infections
- AI-driven triage and resource allocation for ECMO in critical care
Selected Articles
1. Metagenomic next-generation sequencing on treatment strategies and prognosis of patients with lower respiratory tract infections: A systematic review and meta-analysis.
Across 12 studies, mNGS use in LRTIs increased antibiotic adjustments (OR 2.47) and was associated with lower in-hospital mortality (OR 0.49), with consistent benefit in adults, severe LRTI, and BALF-only testing; hospital length of stay was unchanged. Findings support incorporating mNGS into stewardship pathways for severe LRTIs.
Impact: This meta-analysis links pathogen-agnostic sequencing to both therapeutic decision changes and improved survival, providing rare outcome-level evidence in respiratory infection diagnostics.
Clinical Implications: Consider mNGS in severe or diagnostically uncertain LRTIs—especially using BALF—to guide targeted therapy and antimicrobial stewardship; integrate results into rapid MDT review to translate diagnostic gain into timely therapy changes.
Key Findings
- mNGS increased antibiotic change rates in LRTIs (OR 2.47; 95% CI 1.42–4.28).
- mNGS use was associated with reduced in-hospital mortality (OR 0.49; 95% CI 0.36–0.67).
- No significant effect on length of hospital stay (mean difference −1.79 days; 95% CI −5.20 to 1.63).
- Benefits were consistent in adults, severe LRTIs, and in studies using BALF-only testing.
Methodological Strengths
- Prospero-registered protocol with comprehensive database search and predefined outcomes.
- Formal risk-of-bias assessment for RCTs and observational studies; random-effects meta-analysis with subgroup analyses.
Limitations
- Pooled evidence largely observational with heterogeneity in populations and mNGS platforms.
- Potential publication bias and limited RCT data; lack of standardized turnaround times.
Future Directions: Prospective multicenter RCTs integrating mNGS into care pathways (with predefined turnaround and action protocols) to confirm survival benefit and cost-effectiveness.
OBJECTIVES: Controversy exists regarding the benefits of metagenomic next-generation sequencing (mNGS) in lower respiratory tract infections (LRTIs). We assessed the impact of mNGS on the treatment and prognosis of LRTI patients through a systematic review and meta-analysis. METHODS: A literature search was conducted in PubMed, Embase, and CENTRAL databases up to 19 February 2024. Studies investigating the clinical value of mNGS in patients with LRTIs were included. The Risk-of-Bias Tool for randomized controlled trials and the Newcastle-Ottawa scale for observational studies were used to assess risk of bias. Antibiotic change rates and prognostic outcomes were evaluated using random-effects analyses with 95% confidence intervals (CIs). This study is registered with PROSPERO, CRD42024509738. RESULTS: Twelve studies were included in the meta-analysis. The use of mNGS was associated with a higher rate of antibiotic change (odds ratio, 2.47; 95% CI, 1.42-4.28; P < 0.01). Consistent findings were observed in adults, patients with severe LRTIs, and in those who underwent mNGS testing exclusively on bronchoalveolar lavage fluid. We also observed a reduction in in-hospital mortality (odds ratio, 0.49; 95% CI, 0.36-0.67; P < 0.01), though no significant impact on length of hospital stay was observed (mean difference, -1.79; 95% CI, -5.20 -1.63; P = 0.31). CONCLUSIONS: This meta-analysis indicates that the application of mNGS may lead to changes in antibiotic prescriptions for patients with LRTIs, and might reduce the risk of mortality. However, large-scale randomized controlled clinical trials are urgently needed to validate the findings of this study.
2. Surgical and Bronchoscopic Lung Volume Reduction for Severe Emphysema: A Systematic Review and Network Meta-analysis.
Across 25 RCTs (4,283 patients), LVRS achieved the largest functional gains (FEV1, 6MWD, dyspnea) but increased mid-term mortality (RR 3.26). EBV and endobronchial coils improved outcomes but raised pneumothorax risk, with EBV slightly increasing mortality (RR 2.06). Results support individualized selection balancing efficacy and risk.
Impact: This network meta-analysis synthesizes head-to-head and indirect evidence across all major LVR modalities, guiding procedure selection in advanced emphysema where comparative data were fragmented.
Clinical Implications: LVRS should be reserved for carefully selected patients who can tolerate higher early mortality risk in exchange for maximal functional gains; EBV/EBC are alternatives where fissure anatomy and collateral ventilation permit, with vigilant pneumothorax monitoring and shared decision-making.
Key Findings
- Included 25 RCTs (n=4,283) comparing seven LVR modalities versus standard care.
- LVRS provided the largest improvements in FEV1, 6MWD, and symptoms but increased mid-term mortality (RR 3.26; 95% CrI 1.98–6.21).
- EBV improved function but increased pneumothorax and slightly increased mid-term mortality (RR 2.06; 95% CrI 1.07–4.36).
- Endobronchial coils (EBC) showed efficacy but with pneumothorax risk; long-term survival data for BLVR remain limited.
Methodological Strengths
- Network meta-analysis across RCTs enabling indirect comparisons among multiple interventions.
- Comprehensive multi-database search with standardized outcome synthesis.
Limitations
- Mid-term (≤6 months) mortality focus limits inference on long-term survival.
- Heterogeneity in patient selection, fissure completeness, and procedural expertise.
Future Directions: Head-to-head trials and patient-level network meta-analyses with standardized pneumothorax management and long-term survival endpoints to refine modality selection.
BACKGROUND: Along with lung volume reduction surgery (LVRS), bronchoscopic lung volume reduction is a treatment option for end-stage emphysema. However, comparisons among interventions remain insufficient. METHODS: We searched on PubMed, CENTRAL, Embase, and Web of Science. We included randomized controlled trials with outcomes measuring mid-term mortality within 6 months, changes in forced expiratory volume in one second (FEV RESULTS: Twenty-five randomized controlled trials involving 4,283 patients were included, identifying seven types of procedures and standard of care. Mid-term mortality increased in LVRS and endobronchial valve (EBV) (LVRS, risk ratio [RR] 3.26, 95% CrI 1.98-6.21, low certainty; EBV, RR 2.06 95% CrI 1.07-4.36, moderate certainty). LVRS showed the largest improvements: change in FEV CONCLUSION: LVRS offers high efficacies but is accompanied by increased mid-term mortality. EBV and EBC also showed effectiveness; however, they increased pneumothorax, and EBV slightly increased mortality. For accurate assessment, long-term survival data of BLVR are needed.
3. Multi-modal prediction of extracorporeal support-a resource intensive therapy, utilizing a large national database.
Using N3C data (101,400 patients), a hierarchical deep-learning model continuously predicted ECMO utilization up to 96 hours before initiation and outperformed multiple traditional machine-learning baselines at all horizons. Model interpretability highlighted dynamic feature importance across clinical trajectories.
Impact: Provides an actionable, continuously updating triage tool for ECMO—a scarce critical care resource—linking multimodal EHR signals to timely decisions.
Clinical Implications: If prospectively validated, PreEMPT-ECMO could trigger earlier consultation, transfer, or cannulation planning, standardize referral thresholds, and reduce delays for refractory respiratory failure.
Key Findings
- Developed a hierarchical deep-learning model integrating static and multi-granularity time series features from N3C.
- Included 101,400 patients with 1,298 (1.28%) receiving ECMO support.
- Outperformed Logistic Regression, SVM, Random Forest, and XGBoost in accuracy and precision from 0 to 96 hours before ECMO initiation.
- Interpretability analyses revealed evolving feature contributions across patients’ clinical courses.
Methodological Strengths
- Very large, multicenter dataset with multimodal static and time-series features reflecting real-world trajectories.
- Benchmarking against multiple machine-learning baselines and performance assessed across multiple lead times.
Limitations
- Retrospective model development with potential confounding and dataset shift.
- Developed in COVID-era cohorts; generalizability to non-COVID refractory respiratory failure requires prospective validation.
Future Directions: Prospective, multi-center silent deployment with human-in-the-loop evaluation, fairness auditing, and external validation on non-COVID ARDS to assess clinical impact and safety.
OBJECTIVE: Extracorporeal membrane oxygenation (ECMO) is among the most resource-intensive therapies in critical care. The COVID-19 pandemic highlighted the lack of ECMO resource allocation tools. We aimed to develop a continuous ECMO risk prediction model to enhance patient triage and resource allocation. MATERIAL AND METHODS: We leveraged multimodal data from the National COVID Cohort Collaborative (N3C) to develop a hierarchical deep learning model, labeled "PreEMPT-ECMO" (Prediction, Early Monitoring, and Proactive Triage for ECMO) which integrates static and multi-granularity time series features to generate continuous predictions of ECMO utilization. Model performance was assessed across time points ranging from 0 to 96 hours prior to ECMO initiation, using both accuracy and precision metrics. RESULTS: Between January 2020 and May 2023, 101 400 patients were included, with 1298 (1.28%) supported on ECMO. PreEMPT-ECMO outperformed established predictive models, including Logistic Regression, Support Vector Machine, Random Forest, and Extreme Gradient Boosting Tree, in both accuracy and precision at all time points. Model interpretation analysis also highlighted variations in feature contributions through each patient's clinical course. DISCUSSION AND CONCLUSIONS: We developed a hierarchical model for continuous ECMO use prediction, utilizing a large multicenter dataset incorporating both static and time series variables of various granularities. This novel approach reflects the nuanced decision-making process inherent in ECMO initiation and has the potential to be used as an early alert tool to guide patient triage and ECMO resource allocation. Future directions include prospective validation and generalizability on non-COVID-19 refractory respiratory failure, aiming to improve patient outcomes.