Abstract Rationale Reducing extubation failure in high-risk adults is fundamental to ICU practice. HFNC and NIV are used prophylactically after planned extubation, but the comparative effect of modern humidified NIV vs HFNC on short-term re-intubation remains uncertain. Objectives To estimate the effect of prophylactic humidified NIV vs HFNC on 7-day re-intubation after planned extubation in high-risk adult ICU patients. Methods We included RCTs of adults undergoing planned extubation that compared immediate humidified NIV (protocolized ≥24-48 h) with heated-humidified HFNC and reported 7-day re-intubation. Trials using alternating or combination strategies, non-prophylactic indications, pediatrics, or phenotype-restricted cohorts (e.g., cardiac surgery only, COPD-only hypercapnia) were excluded from the primary analysis. Risk ratios (RR) were pooled with random effects using REML and Hartung-Knapp confidence intervals (primary); DerSimonian-Laird with classic random-effects confidence intervals was a prespecified sensitivity. Heterogeneity was summarized with I². A related 72-hour RCT was summarized separately. Results Two eligible RCTs enrolled high-risk adults (one very-high-risk cohort; one obesity cohort). N = 326 were analyzed (NIV 164; HFNC 162). Trial-level RRs for 7-day re-intubation were 0.59 (95% CI 0.37-0.93) and 0.71 (0.42-1.20). The primary pooled estimate was RR 0.64 (95% CI 0.20-2.07), I² 0%, heterogeneity p = 0.60. The DL sensitivity yielded RR 0.64 (95% CI 0.45-0.90), p = 0.01. Across pooled trials, re-intubation occurred in 38/164 (23.2%) with NIV vs 59/162 (36.4%) with HFNC, absolute risk difference -13.2% (approximate NNT 8). The 72-hour RCT (shorter NIV dose, different rescue policy) showed RR 0.84 (95% CI 0.62-1.15) and was not pooled with the 7-day endpoint. Conclusions Within a strict and clinically coherent framework for high-risk adults after planned extubation, humidified NIV is associated with fewer 7-day re-intubations than HFNC. The point estimate is stable across models with no observed heterogeneity. Precision is limited with Hartung-Knapp small-k inference, while the DL sensitivity suggests a statistically significant reduction. These findings support protocolized humidified NIV pathways for high-risk extubations and justify a harmonized confirmatory RCT with standardized endpoint windows and rescue policies. This abstract is funded by: None
Abstract Background Artificial intelligence (AI) and machine-learning (ML) frameworks are redefining the understanding of sepsis by identifying distinct biological endotypes and improving early risk stratification. However, heterogeneity in modeling approaches and outcome definitions has impeded direct clinical translation. This systematic review and meta-analysis synthesized quantitative evidence on diagnostic performance of ML models and qualitatively examined data-driven sepsis phenotyping frameworks across critical-care populations. Methods Following PRISMA-MA 2020 methodology, eight studies published between 2015 and 2025 were included: six ML prediction studies reporting model discrimination metrics and two phenotyping investigations. Random-effects inverse-variance meta-analyses (DerSimonian-Laird estimator) were performed on logit-transformed AUROC, sensitivity, and specificity. Heterogeneity was quantified with I² and τ². Phenotype-only studies were narratively summarized to contextualize mechanistic insights beyond predictive performance. Results Six ML studies collectively analyzed >12 000 patients, with average 58 % male predominance, from multiple international ICU databases. The pooled sensitivity 0.86 [0.75-0.93] and specificity 0.86 [0.71-0.94]; heterogeneity was substantial (I² > 95 %). Despite methodological variation, ensemble gradient boosting, recurrent neural networks, and hybrid rule-based models consistently outperformed traditional severity indices such as SOFA and SAPS II. Model interpretability analyses frequently highlighted lactate kinetics, MAP trends, and inflammatory biomarkers as dominant features. Unsupervised and semi-supervised frameworks—including k-means, Two-Step Cluster, and finite-mixture modeling—consistently delineated 2-5 reproducible sepsis phenotypes differentiated by metabolic signatures, inflammatory profiles, and organ-dysfunction trajectories. Low-lysophospholipid metabolic states and NK-cell-exhaustion transcriptomic patterns were repeatedly associated with excess mortality. Predictive tools such as SORP (AUC 0.95) and SepxFindeR provided accurate early-warning capability and biological interpretability. Individualized treatment simulations, including Dynamic Treatment Regimens, suggested phenotype-specific responses to fluid and vasopressor therapy. Both phenotype studies confirmed internal or partial external validation, supporting emerging translational robustness. Conclusions AI-driven clustering and predictive modeling collectively redefine septic shock as a continuum of biologically and clinically distinct endotypes. Integrative ML pipelines that combine clinical, metabolic, and transcriptomic data enable precise early warning, mechanistic biomarker discovery, and personalized resuscitation strategies. While pooled model discrimination remains high, substantial heterogeneity underscores the need for standardized pipelines, transparent calibration reporting, and multicenter real-time validation to transition ML-based sepsis phenotyping from computational insight to precision critical-care practice. This abstract is funded by: None
Introduction:Metabolic dysfunction-associated steatotic liver disease (MASLD) lacks treatment options, with few evidence-based choices beyond lifestyle modification. This is the first clinical study to assess the efficacy of an organic liquid oral nutraceutical composed of turmeric extract, dandelion powder, milk thistle extract, and ginger powder in MASLD. Methods:This was a randomized, double-blind, placebo-controlled, preliminary phase 2 exploratory trial conducted at a single center (Sanjivani Super Speciality Hospitals Pvt. Ltd., Gujarat, India) between September 7, 2021 and November 25, 2021. Individuals aged 18-70 years and diagnosed with fatty liver (grades 1-3) were eligible to participate. Participants were randomized 1:1 to receive the test product or placebo, which were both provided as a 60 mL oral solution to be consumed twice daily with a meal for 60 days. The primary endpoint was proportion of participants with improvement in ultrasonography-based fatty liver grading from baseline to Day 60. Results:Fifteen patients each were enrolled in the test product and placebo groups. The mean [standard deviation (SD)] age was 40.6 (10.8) years and 83.3% were male. A significantly higher proportion of patients in the nutraceutical product group had fatty liver grade improvement at Day 60 vs. placebo [60.0% (9/15) vs. 13.3% (2/15); p = 0.008]. Mean (SD) change in fatty liver grading was -0.60 (0.51) and -0.13 (0.35) for test product vs. placebo (p = 0.007). Adverse events were mild with similar incidence in both groups. The results are limited by the small sample size, single-center design, and short study duration. Discussion:This study provides preliminary evidence that this organic nutraceutical may improve steatosis grade in individuals with MASLD. Clinical trial registration:Clinical Trial Registry of India (CTRI/2021/08/035359).
Abstract Rationale Center-based pulmonary rehabilitation (PR) improves outcomes, but access and adherence are limited. Whether home-based PR preserves functional benefit while improving completion across mixed chronic respiratory diseases is clinically important. Objectives To compare home-based with center-based PR for (1) change in 6-minute walk distance (6MWD) using a prespecified equivalence framework and (2) program completion. Methods Randomized trials directly comparing home-based (including tele-supported) and center-based PR were pooled. The primary endpoint was end-of-program change in 6MWD. Equivalence margin: ±30 m. Random-effects (REML) models estimated pooled mean differences (MD) with 95% CIs and prediction intervals (PI); equivalence was tested with two one-sided tests (TOST). Hartung-Knapp (HK) adjustment was used as a conservative sensitivity model. Completion was synthesized as risk ratio (RR; Mantel-Haenszel). We report pooled control completion risk, absolute risk difference (ARD), and number-needed-to-treat (NNT). Heterogeneity was summarized with I². Results Three RCTs were included (k = 3).6MWD: Pooled MD (home − center) −0.42 m (95% CI − 21.74 to 20.89; I²=55.2%); PI − 35.10 to 34.26 m. TOST supported equivalence within ±30 m (p = 0.0033). With HK, the MD was unchanged but the CI widened to − 47.78 to 46.94 m, yielding an inconclusive equivalence inference under that conservative model.Completion: Across 396 participants (193 home; 203 center), fixed-effect RR was 1.45 (95% CI 1.28-1.65). Random-effects with HK: RR 1.46 (95% CI 0.72-2.94); I²=86.5%; PI ∼0.80-2.70. The pooled control completion risk was 58.6%. Using the point estimate, ARD was +26.9% and NNT ≈ 3.7, acknowledging the wide random-effects/HK uncertainty. Conclusions In adults with chronic respiratory diseases, home-based PR delivers 6MWD gains equivalent to center-based PR using a ± 30 m margin under standard random-effects modeling; HK widens uncertainty but does not suggest inferiority. Completion is likely higher with home-based PR, though the magnitude varies across implementations (high heterogeneity). Programs struggling with uptake can consider structured home pathways to maintain functional benefit while improving adherence. Future trials should standardize completion definitions, report responder outcomes, and detail implementation components (e.g., coaching intensity, remote monitoring) to explain heterogeneity and guide scale-up. This abstract is funded by: None
Abstract Background Sepsis-associated acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) represent the most lethal manifestations of dysregulated systemic inflammation. Despite decades of research, no approved pharmacologic therapies directly target the inflammatory and epithelial-endothelial injury pathways that drive sepsis-related respiratory failure. Glucagon-like peptide-1 receptor agonists (GLP-1RAs), originally developed for diabetes, exert potent anti-inflammatory, antioxidative, and cytoprotective actions that may interrupt these injury cascades. This systematic review synthesizes experimental evidence evaluating the role of GLP-1RAs in preventing or mitigating sepsis-associated ALI. Methods Following PRISMA 2020 guidelines, PubMed, Embase, and Scopus were searched through October 2025 for preclinical or clinical studies assessing GLP-1 analogs (liraglutide, dulaglutide, semaglutide, exenatide) in sepsis or pneumonia-related lung injury. Eligible studies examined outcomes related to pulmonary inflammation, surfactant function, oxidative stress, vascular injury, or clinical respiratory endpoints. Mechanistic pathways were categorized into immune modulation, oxidative stress, apoptosis, and epithelial barrier preservation. Results Five studies met inclusion criteria (four translational, one large human cohort). Across murine models of endotoxemia, cecal slurry-induced sepsis, and pneumonia-sepsis, GLP-1RA treatment consistently reduced pro-inflammatory cytokines (TNF-α, IL-6, IL-1β) by 40-70%, suppressed NLRP3/STAT3 signaling, and inhibited apoptosis and autophagy (↓ markers: caspase-3, Bax/Bcl-2, Beclin-1, LC3). Liraglutide and dulaglutide improved surfactant protein (SP-A/SP-B) synthesis, preserved alveolar-capillary integrity, decreased lung edema, and enhanced survival by 25-50%. Mechanistically, GLP-1 receptor activation attenuated oxidative stress and restored endothelial function through downregulation of iNOS and reactive nitrogen species. The only clinical study demonstrated a ∼40% reduction in incident pneumonia and severe sepsis among GLP-1RA users, suggesting potential translational relevance. Conclusions Evidence from translational and observational studies indicates that GLP-1 receptor agonists mitigate sepsis-associated ALI by attenuating inflammation, oxidative stress, and apoptosis, while promoting surfactant homeostasis and vascular barrier function. GLP1-RAs appear to stabilize both vascular and epithelial compartments of the lung microenvironment, thereby improving host resilience to infection-induced injury. These findings identify GLP-1 signaling as a promising immunometabolic therapeutic target for sepsis-related respiratory failure. Future prospective randomized trials are warranted to define optimal dosing, timing, and patient selection for GLP-1RA use in this subgroup. This abstract is funded by: None
