BACKGROUND: Etomidate, ketamine, and propofol are all used as induction agents for emergency tracheal intubation in critically ill adults but it remains uncertain which agent should be preferable. METHODS: We searched ME...BACKGROUND: Etomidate, ketamine, and propofol are all used as induction agents for emergency tracheal intubation in critically ill adults but it remains uncertain which agent should be preferable. METHODS: We searched MEDLINE and Embase (inception to December 2025) for randomized controlled trials comparing etomidate, ketamine, propofol, or ketamine-propofol combination (ketofol) for emergency or rapid sequence intubation in critically ill adults. We performed random-effects network meta-analysis using the frequentist framework. The primary outcome was short-term mortality (28-30 day, or ICU/in-hospital mortality when unavailable). Secondary outcomes included cardiovascular collapse, post-induction hypotension, vasopressor use, first-pass intubation success, and peri-intubation cardiac arrest. Certainty of evidence was assessed using the CINeMA framework. RESULTS: Nine trials (4,672 patients, four treatments) were included. Ketamine and etomidate probably result in similar mortality (OR 0.96, 95% CI 0.80-1.16; [Formula: see text] = 30%; moderate certainty). Evidence for other mortality comparisons was very uncertain: ketamine vs propofol (OR 1.53, 0.80-2.93; 1 trial; low certainty) and etomidate vs propofol (OR 0.63, 0.32-1.24; indirect only; very low certainty). Compared with etomidate, ketamine probably increases cardiovascular collapse (OR 1.44, 1.20-1.71; moderate certainty) and may increase post-induction hypotension (OR 1.34, 1.07-1.68; low certainty) and peri-intubation vasopressor use (OR 1.45, 1.21-1.74; low certainty). There was probably little or no difference in first-pass intubation success or cardiac arrest. CONCLUSIONS: Etomidate and ketamine probably result in similar mortality, but confidence intervals are compatible with clinically important differences in either direction-ketamine probably causes more peri-intubation hemodynamic instability. Beyond one trial, no randomized evidence exists for propofol in emergency intubation of critically ill adults.
BACKGROUND: Acute kidney injury (AKI) is a frequent, severe complication in the intensive care units (ICU). Existing machine learning models are typically inflexible, classification-based (i.e., predicting AKI occurrence...BACKGROUND: Acute kidney injury (AKI) is a frequent, severe complication in the intensive care units (ICU). Existing machine learning models are typically inflexible, classification-based (i.e., predicting AKI occurrence as yes/no), and of limited clinical utility. This study proposes and externally validates the first multi-step, multivariate distributional regression model that directly predicts future distributions of serum creatinine (sCr) and urine output across multiple time horizons, thereby enhancing AKI risk stratification and personalized clinical decision support. METHODS: The model was developed using a training cohort of 4,118 adult ICU stays from the MIMIC-IV dataset and externally validated on four independent, diverse cohorts: MIMIC-IV (N=3,838), UZGent (N=4,442), eICU (N=10,760), and AmsterdamUMC (N=6,129). The model used clinical data to generate multivariate predictive distributions hourly for urine output and sCr (up to 48 hours ahead). Predictors included demographics, vital signs, laboratory results, medications, and recent urine output, with time-varying variables summarized over the preceding 72 hours (recent value, slope, minimum, maximum, variability). Performance was evaluated by comparing our predictive distributions with state-of-the-art tree-based classifiers for 24-hour ahead prediction of KDIGO stages 1-3 AKI and persistent stage 3 AKI. RESULTS: Across all external cohorts, the distributional regression model demonstrated high discrimination (mean AUC-PR 0.774 for all stages) and excellent calibration, consistently outperforming the benchmark classifiers. By jointly predicting sCr and urine output distributions, a single model successfully enables flexible risk stratification across all stages, capturing AKI onset and persistence, and allowing changes to stage definitions. CONCLUSION: This multi-step, multivariate distributional regression model is a reliable, more flexible, transparent, and clinically interpretable approach for AKI prediction compared to traditional classification methods. It represents a necessary step toward bedside implementation of predictive models for personalized AKI management in the ICU.
BACKGROUND: The intracranial pressure (ICP) pulse waveform reflects intracranial compliance. The pulse shape index (PSI), an artificial intelligence (AI)-based metric ranging from 1 (normal) to 4 (disturbed), quantifies...BACKGROUND: The intracranial pressure (ICP) pulse waveform reflects intracranial compliance. The pulse shape index (PSI), an artificial intelligence (AI)-based metric ranging from 1 (normal) to 4 (disturbed), quantifies morphological waveform pathologies. Early findings in smaller cohorts indicate that elevated PSI is associated with mass lesions, aging, higher ICP, and worse outcome. This study examined how PSI relates to other markers of intracranial compliance, the risk of ICP crisis, and clinical outcome in a large traumatic brain injury (TBI) cohort. METHODS: This retrospective study included 321 TBI patients with ≥ 12 h of ICP/PSI monitoring. PSI was analysed in relation to ICP, ICP pulse amplitude (AmpICP), the moving correlation coefficient between mean ICP and ICP amplitude (RAP index), and arterial blood pressure (ABP) using generalised additive models (GAMs). Linear mixed-effects models assessed whether PSI during preceding hours predicted later ICP elevations (e.g., ICP > 20/22 mmHg). The prognostic value of PSI was tested using univariable Mann-Whitney U test and multivariable (after adjustment for age, Glasgow Coma Scale, ICP, CPP, and PRx) logistic regression for mortality (Glasgow Outcome Scale = 1) and favourable outcome (Glasgow Outcome Scale = 4-5). RESULTS: PSI increased with higher ICP, AmpICP, RAP, and ABP. PSI during the preceding hour predicted increased ICP burden above 20 mmHg (marginal R ~30%). PSI was lower in survivors and in patients with favourable outcomes. However, in multivariable logistic regressions, PSI was not independently associated with mortality or favourable outcome. CONCLUSIONS: PSI was linked to established indicators of intracranial compliance and provided early warning of impending intracranial hypertension. While PSI showed univariable associations with outcome, these did not remain in multivariable models. Retrospectively, PSI does not appear to have independent prognostic value, but rather a complementary value. Prospectively, it may serve as a dynamic marker of deteriorating intracranial compliance and an early signal for future ICP crisis.
BACKGROUND: The timing of declaration of death by neurologic criteria (DNC) after acute brain injury varies in clinical practice. However, large-scale data describing etiology- and age-specific timing of DNC evaluation a...BACKGROUND: The timing of declaration of death by neurologic criteria (DNC) after acute brain injury varies in clinical practice. However, large-scale data describing etiology- and age-specific timing of DNC evaluation and declaration are limited. This study aimed to characterize these timing patterns in a cohort of patients with DNC. METHODS: Retrospective cohort study using registry data from German hospitals reporting cases of guideline-confirmed DNC to the national organ procurement organization. Included were adults (≥ 18 years) who were declared DNC between April 1, 2020, and August 31, 2025. Brain injury was categorized by ICD-10 as traumatic brain injury (TBI), intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), acute ischemic stroke (AIS), hypoxic-ischemic encephalopathy (HIE). The primary outcome was time from initiation of mechanical ventilation to DNC declaration; time to first clinical examination fulfilling criteria for DNC was analyzed as a secondary outcome. Time-to-event distributions were summarized using cumulative incidence estimates, and associations of timing with etiology, age, and selected clinical variables were explored using random survival forests. RESULTS: Among 6,701 patients with DNC from 716 hospitals, 6,397 had complete timing data (median age, 59 years [IQR, 46-69]; 54% male). Median time to DNC differed across etiologies (p < 0.001): 55.5 h for ICH, 58.4 h for TBI, 65.9 h for SAH, 89.9 h for HIE and 96.8 h for AIS. By 72 h of ventilation, cumulative incidence of DNC was 64.5% (95%CI, 61.3-67.5) in TBI, 62.4% (95%CI, 59.9-64.7) in ICH, 53.7% (95%CI, 50.9-56.4) in SAH, 36.7% (95%CI, 34.4-38.9) in HIE and 31.9% (95%CI, 28.3-35.4) in AIS. By 120 h, cumulative incidence reached 83.9% (95%CI, 81.3-86.1) in TBI, 81.5% (95%CI, 79.5-83.4) in ICH, 69.0% (95%CI, 66.4-71.5) in SAH, 67.3% (95%CI, 65.0-69.4) in HIE and 65.7% (95%CI, 61.9-69.1) in AIS. Older age was associated with shorter time to DNC in TBI and ICH, whereas associations were minimal in AIS and HIE. CONCLUSIONS: The timing of DNC evaluation and declaration varies substantially across etiologies and age groups. These findings may help to better understand temporal patterns of DNC determination among patients with DNC in clinical practice.
The Global consensus definition of acute respiratory distress syndrome (ARDS) broadened the syndrome to include patients on high-flow nasal cannula and hypoxia as defined by the ratio of the saturation of oxygen to fract...The Global consensus definition of acute respiratory distress syndrome (ARDS) broadened the syndrome to include patients on high-flow nasal cannula and hypoxia as defined by the ratio of the saturation of oxygen to fraction of inhaled oxygen (SFR). We sought to compare the incidence and outcomes of ARDS under the 2012 Berlin versus 2023 Global definitions, and to examine the relationship between SFR-derived categories of severity and mortality, in a prospective cohort of critically ill patients with sepsis. Of the 950 included patients, 466 (49%) met criteria for ARDS under the Global definition and 427 (45%) met criteria for ARDS under the Berlin definition during the 6-day follow-up period. Among patients with ARDS, the Global definition allowed for ARDS qualification a median of 3.0 h earlier than the Berlin definition. Mortality was comparable between the Global and Berlin definitions at onset but substantially lower for patients who never went onto meet the Berlin definition. SFR was predictive of 30-day mortality and exhibited moderate correlation with the ratio of partial pressure of oxygen to fraction of inhaled oxygen (PFR). Our work establishes an increased incidence and modestly decreased time to diagnosis of ARDS under the Global definition and supports the prognostic validity of SFR.
Rosiewicz KS, Anft M, Stervbo U
… +17 more, Skrzypczyk S, Kaliszczyk S, Koos B, Rump K, Nowak H, Unterberg M, Westhoff TH, Brenner T, Trilling M, Schmueck-Henneresse M, Wolk K, Sabat R, Gregorius J, Wappler F, Zarbock A, Adamzik M, Babel N
BACKGROUND: Sepsis is increasingly recognized as a highly dynamic immunological disorder in which hyperinflammation and anti-inflammatory processes occur simultaneously. The clinical phenotype depends on which arm predom...BACKGROUND: Sepsis is increasingly recognized as a highly dynamic immunological disorder in which hyperinflammation and anti-inflammatory processes occur simultaneously. The clinical phenotype depends on which arm predominates at a given time, resulting in an early phase that is typically dominated by hyperinflammation and a subsequent phase characterized by hypoinflammation, also referred to as immunoparalysis (IP). Epstein-Barr virus (EBV) reactivation has been associated with an immunosuppressive status. However, its interaction with IP and resulting immune phenotypes remains poorly defined so far. In this current study, we investigated the temporal dynamics of EBV reactivation and IP status, and assessed their impact on immune signatures and mortality in sepsis. METHODS: In this retrospective cohort of 124 intensive care unit (ICU) patients with sepsis, we performed analysis of EBV load by qPCR and analyzed the inflammatory stage by quantifying HLA-DR molecules on monocytes (mHLA-DR) using flow cytometry. Patients with < 5,000 mHLA-DR/monocyte were classified positive for immunoparalysis (IP+). Cytokine profiles and vital sign were analyzed in parallel. Patients were assigned to four sepsis groups based on EBV/IP status (EBV- IP-, EBV + IP-, EBV- IP+, EBV + IP+). Time-dependent Cox models (start-stop structure) were used to estimate hazard ratios (HR) for mortality, adjusted for age, sex and Sequential Organ Failure Assessment (SOFA) score. Cytokines and clinical markers were compared using Kruskal-Wallis and rank-based analyses. RESULTS: EBV positivity was associated with higher hazard of death (HR 3.30, 95% CI 1.24-8.81, p = 0.0009), while IP showed a similar but nonsignificant trend (HR 2.14, 95% CI 0.93-4.90, p = 0.073). The combined EBV + IP+ group exhibited the highest mortality (HR 7.23, 95% CI 2.24-23.3, p = 0.0009). This group also showed the strongest cytokine activation (IL-6, IL-8, IL-10, IL-17 A, IL-18, MCP-1) and lowest mHLA-DR expression, indicating a mixed hyperinflammatory and immunosuppressed phenotype. In contrast, EBV- IP- patients displayed the most immunocompetent baseline profile. CONCLUSION: Our preliminary findings suggest that EBV reactivation superimposed on immunoparalysis is associated with a harmful sepsis endotype combining excessive cytokine activity with impaired monocyte function. Further studies are needed to evaluate if the dynamic monitoring of EBV DNA and mHLA-DR expression may enable early identification of patients at highest risk and guide targeted immunomodulatory or antiviral interventions.
Both venous excess ultrasound (VExUS) and central venous pressure (CVP) can assess venous congestion from different perspectives. CVP provides continuous and simple monitoring of the risk of congestion and is easily repe...Both venous excess ultrasound (VExUS) and central venous pressure (CVP) can assess venous congestion from different perspectives. CVP provides continuous and simple monitoring of the risk of congestion and is easily repeatable in patients with a central venous catheter. In contrast, VExUS offers an intermittent, organ-level evaluation of established congestion and may help identify the venous waterfall phenomenon. Furthermore, assessment of pulmonary congestion should not be overlooked.
Gakuba C, Denis L, Chabouh G
… +13 more, Gauberti M, Moyer JD, Gavet T, Bourgès J, Malczuk J, Briant A, Morello R, Laquay N, Bodard S, Chavignon A, Hingot V, Vivien D, Couture O
BACKGROUND: Multimorbidity is prevalent among critically ill patients with sepsis yet remains underreported in critical care randomised controlled trials (RCTs). Inadequate reporting limits the generalisability of findin...BACKGROUND: Multimorbidity is prevalent among critically ill patients with sepsis yet remains underreported in critical care randomised controlled trials (RCTs). Inadequate reporting limits the generalisability of findings and the ability to understand whether chronic disease burden modifies treatment effects. We aimed to systematically map and evaluate how comorbidities and multimorbidity are represented, reported, and analysed in RCTs involving Intensive Care (ICU) patients with sepsis or septic shock. METHODS: We searched MEDLINE, Embase, CENTRAL, LILACS, Web of Science, medRxiv and clinical trial registries for RCTs between January 1992 and August 2025 for trials enrolling adult ICU patients with sepsis or septic shock. Two reviewers independently screened studies and extracted data using a predefined protocol registered on PROSPERO (CRD42024510506). Trials were categorised according to whether baseline comorbidity information was reported. Reported comorbidities were mapped to a Delphi-derived classification framework. RESULTS: From 15,830 records, 591 RCTs met inclusion criteria. Of these, 209 trials (35.4%) reported baseline comorbidity data, enrolling a total of 48,429 patients (median 91 per trial). Trials reporting comorbidity information differed systematically from those that did not: they were larger, more likely to have publicly available protocols, more frequently used mortality as a primary outcome, and more often demonstrated low risk of bias in several methodological domains. Reporting of comorbidities increased over time (ρ = 0.80, p < 0.001), with the odds of reporting baseline comorbidity data increasing by approximately 8% per year (OR 1.08, 95% CI 1.06-1.11). Among trials reporting comorbidity data, the most frequently reported conditions were diabetes (86.1% of trials), hypertension (65.1%), chronic kidney disease (56%), and cancer (53.1%). Despite the high prevalence of comorbidities, only four trials (1.9%) explicitly reported multimorbidity and just 12 trials (5.7%) used a structured framework such as the Charlson Comorbidity Index. Nearly 80% of trials excluded participants based on at least one comorbidity. CONCLUSIONS: Baseline comorbidity data are absent from most sepsis RCTs, and trials that report such information differ systematically from those that do not. Standardised frameworks and transparent reporting of comorbidities and multimorbidity are needed to improve representativeness, enable subgroup analyses, and enhance the external validity of sepsis research.
BACKGROUND: The Sepsis-3 criteria operationalized organ dysfunction using the original Sequential Organ Failure Assessment (SOFA-1) score, which was updated to SOFA-2 in October 2025 to align with modern intensive care u...BACKGROUND: The Sepsis-3 criteria operationalized organ dysfunction using the original Sequential Organ Failure Assessment (SOFA-1) score, which was updated to SOFA-2 in October 2025 to align with modern intensive care unit (ICU) practices. However, the impact of adopting SOFA-2 for sepsis detection under the Sepsis-3 criteria has not yet been evaluated. METHODS: We conducted a retrospective multicenter cohort study using three large-scale ICU databases from the United States and the Netherlands. Adult patients with suspected infection within 72 h of ICU admission were included. Sepsis was independently identified according to Sepsis-3 criteria, utilizing either the SOFA-1 or SOFA-2 score. We systematically compared diagnostic concordance, the timeliness of sepsis detection, clinical outcomes and predictive performance of prognostic models between the two scoring systems. The primary outcome was ICU mortality, while secondary outcomes included hospital mortality and 28-day survival. RESULTS: The study cohort comprised 74,615 adult patients with suspected infection. The diagnostic concordance of sepsis between SOFA-1 and SOFA-2 reached 89.62%. However, SOFA-1 and SOFA-2 uniquely identified an additional 3.54% and 6.84% of patients as having sepsis, respectively. The diagnostic discrepancies were primarily attributable to updates in respiratory and renal scoring criteria. ICU mortality was highest among the Concordant Positive group (15.63%). Notably, both discordant groups exhibited substantial mortality (SOFA-1 Only: 8.31%; SOFA-2 Only: 9.23%), both of which were significantly higher than those of patients not classified as having sepsis by either score (6.71%; P = .002 and P = 2.87 × 10). Within the Concordant Positive group, 61.24% of patients were diagnosed simultaneously by both criteria. However, SOFA-2 achieved earlier diagnosis in a greater proportion of cases than SOFA-1 (23.12% vs. 15.64%), despite heterogeneity across different databases. Predictive models derived from the SOFA-2 score demonstrated numerically higher area under the receiver operating characteristic curve (AUROC) values in forecasting ICU mortality than those based on SOFA-1 (0.736 vs. 0.728 in internal cross-validation, P = .386; 0.743 vs. 0.720 in external validation, P = .375). CONCLUSIONS: SOFA-1 and SOFA-2 showed high concordance in sepsis detection, yet each identified distinct patient subgroups with significant mortality. A transitional strategy utilizing both SOFA-1 and SOFA-2 is advised until updated expert-validated sepsis criteria are established.
BACKGROUND: Most patients with out-of-hospital cardiac arrest do not achieve sustained return of spontaneous circulation (ROSC). Resuscitative endovascular balloon occlusion of the aorta (REBOA) may increase blood pressu...BACKGROUND: Most patients with out-of-hospital cardiac arrest do not achieve sustained return of spontaneous circulation (ROSC). Resuscitative endovascular balloon occlusion of the aorta (REBOA) may increase blood pressure proximal to the ballon. If this technique is used during advanced life support (ALS), and occlusion is performed in the thoracic aorta, it may augment aortic pressure and coronary perfusion pressure. We investigated whether prehospital REBOA as an adjunct to ALS increased the rate of ROSC. METHODS: REBOARREST was a pragmatic, parallel-group, multicentre, randomised controlled trial conducted at 12 sites in Norway, Denmark, and Italy. Adult patients (18-80 years) with non-traumatic out-of-hospital cardiac arrest were randomly assigned (1:1) to either a control group that received ALS or to an intervention group that received ALS combined with REBOA as an adjunct. Fulfilment of eligibility criteria was determined by the physician on scene and sealed envelopes were used to allocate patients. The statistician that performed the analyses was blinded for group allocation. The primary outcome was sustained ROSC, defined as lasting ≥ 20 min, assessed in the intention-to-treat population. RESULTS: From June 7, 2021, to June 28, 2025, 200 patients were randomly assigned to the study groups. Due to lack of consent 21 patients dropped out of the trial, hence data from 179 patients are presented, 88 in the intervention group and 91 in the control group. Most patients were male (76%), with median age of 68 years (IQR 58-74). Median time from arrest to randomisation was 33 min (IQR 23-39) in the intervention group and 29 min (IQR 23-38) in the control group. Twenty-five of 88 patients (28%) in the intervention group and 24 of 91 patients (26%) in the control group achieved sustained ROSC (adjusted risk difference 1.8% [-11, 15, 95% CI], p = 0.78). Adverse events were registered in 19 patients. CONCLUSIONS: Among patients with non-traumatic out-of-hospital cardiac arrest, a strategy of prehospital deployment of REBOA as an adjunct to ALS was feasible but did not significantly improve rates of sustained ROSC compared to ALS alone. Deployment of prehospital REBOA is safe and manageable in a two-person team with low procedure time. TRIAL REGISTRATION: Clinicaltrials.gov ID NCT04596514. Registered 22.10.2020.
BACKGROUND: In severe acute kidney injury (AKI), delayed renal replacement therapy (RRT) strategies allow many KDIGO stage-3 patients to avoid dialysis, but excessive postponement in those who ultimately require RRT may...BACKGROUND: In severe acute kidney injury (AKI), delayed renal replacement therapy (RRT) strategies allow many KDIGO stage-3 patients to avoid dialysis, but excessive postponement in those who ultimately require RRT may worsen outcomes. Early physiologically grounded markers to identify patients likely to need RRT are lacking. We evaluated whether combining early glomerular filtration kinetics and timed urinary urea excretion could improve discrimination of subsequent RRT initiation under a delayed strategy. METHODS: TUBSAKI is a prospective bicentric ICU cohort including adults with KDIGO stage-3 AKI managed with a protocolized delayed RRT strategy. Blood and 24-hour urine samples were collected at diagnosis (D0) and day 1 (D1). Glomerular filtration dynamics were assessed using kinetic GFR (kGFR), and timed urinary urea excretion was assessed using UUEI. Discrimination for subsequent RRT was assessed using ROC curves and AUC. A combined logistic model (kGFR D0-D1 + UUEI D1) was internally validated by bootstrap, with sensitivity analyses adjusted for SOFA and KDIGO stage-3 oliguria. RESULTS: Among 110 patients, 31 (28%) required RRT. kGFR D0-D1 showed good discrimination (AUC 0.81 [0.72-0.89]), and UUEI D1 moderate discrimination (AUC 0.74 [0.63-0.82]). The combined model showed an AUC of 0.85 ([0.76-0.91]), optimism-corrected AUC 0.83, and acceptable calibration. Discrimination remained stable after adjustment for SOFA and oliguria. Incremental gain over kGFR alone was modest and not statistically significant. CONCLUSIONS: Early glomerular filtration kinetics and urinary urea excretion were associated with subsequent RRT initiation under a delayed strategy. The incremental clinical value of UUEI remained limited in this cohort, and external validation is required before clinical use.
The glymphatic system facilitates cerebrospinal fluid (CSF)-interstitial fluid exchange and plays a key role in solute clearance and neurophysiological homeostasis. While dysfunction of this system has been shown in trau...The glymphatic system facilitates cerebrospinal fluid (CSF)-interstitial fluid exchange and plays a key role in solute clearance and neurophysiological homeostasis. While dysfunction of this system has been shown in traumatic brain injury, stroke, meningitis, idiopathic normal pressure hydrocephalus and neurodegenerative diseases, direct measurement of glymphatic transport in humans remains elusive. We propose microGLYMPH as a translational, hypothesis-generating framework that combines established clinical cerebral microdialysis with controlled CSF tracer administration via existing clinical access routes, including an external ventricular drain, cisternal access during surgery, or lumbar intrathecal injection when clinically justified. The aim is to obtain time-resolved regional tracer profiles in microdialysate and to interpret these alongside arousal state, intracranial dynamics, and, where available, complementary imaging, thereby providing an indirect measure of CSF-interstitial exchange kinetics and peripheral tracer appearance. We further define the key design, analytical and practical limitations that must be resolved before the approach can extend beyond exploratory use, notably catheter-adjacent effects, blood-brain barrier disruption, drainage practices, and the intrinsically focal nature of microdialysis. microGLYMPH is therefore intended as a staged roadmap for first-in-human feasibility studies and subsequent hypothesis-driven investigations of neurofluid solute transport after acute brain injury.
Bouras M, Pageau M, Gagnon MA
… +14 more, Costerousse O, Demers K, Grenier-Gagnon A, Isaac CJ, Torkomyan TH, Lauzier F, Zarychanski R, Francoeur CL, Gerges P, Abiala G, Moore L, English SW, Turgeon AF, Canadian Traumatic Brain Injury Research Consortium
BACKGROUND: Moderate to severe traumatic brain injury (TBI) is associated with high rates of mortality and long-term disability. Accurate biomarkers are needed to predict longterm neurological outcomes and guide decision...BACKGROUND: Moderate to severe traumatic brain injury (TBI) is associated with high rates of mortality and long-term disability. Accurate biomarkers are needed to predict longterm neurological outcomes and guide decision-making early after TBI. Neurofilament light chain (NfL), a structural protein of neurons, has emerged as a promising candidate, but its association with outcomes in this population remains uncertain. METHODS: We conducted a systematic review and meta-analysis to assess the association between blood or cerebrospinal fluid NfL concentrations and outcomes in adults with moderate to severe TBI. We searched MEDLINE, Embase, Cochrane CENTRAL and Web of Science from inception to October 2025. Eligible studies included cohort studies or randomized controlled trials reporting NfL levels measured during the acute phase and reporting at least one outcome of interest. Our primary outcome was long-term neurological function, defined as the latest available Glasgow Outcome Scale (GOS) or Glasgow Outcome Scale-Extended (GOS-E) score, dichotomized into unfavorable (GOS ≤ 3 or GOS-E ≤ 4) and favorable (GOS > 3 or GOS-E > 4). Mortality, at any time point, was a secondary outcome. Risk of bias was assessed using an adapted scale from the QUADAS-2 tool, and certainty of evidence was evaluated using GRADE criteria. RESULTS: Fourteen studies (2,905 participants) were included, with ten (n = 1,648) contributing to the meta-analysis for our primary outcome. Higher NfL concentrations were associated with unfavorable neurological outcomes, with moderately higher levels in patients with poor outcomes compared with those with favorable outcomes (SMD 0.45, 95% CI 0.33-0.56; I² = 12%). Six studies (n = 483) assessed mortality; higher NfL concentrations were associated with increased mortality (SMD 0.71, 95% CI 0.04-1.39; I² = 82%), with a more consistent association when NfL was measured within 24 h after injury (I² = 0%). The certainty of evidence was graded as very low for both outcomes, reflecting risk of bias and, for mortality, additional inconsistency and imprecision. CONCLUSIONS: Higher NfL concentrations were associated with unfavorable neurological outcomes after moderate-to-severe TBI. The association with mortality was more uncertain and should be interpreted with caution given the substantial heterogeneity across studies. Its incremental prognostic value beyond known predictors remains uncertain. TRIAL REGISTRATION: PROSPERO CRD42022332110, 22 May 2022.
BACKGROUND: The use of extracorporeal membrane oxygenation (ECMO) has expanded for severe respiratory and circulatory failure. Acute kidney injury (AKI) is one of the most frequent complications. Up to 85% of ECMO patien...BACKGROUND: The use of extracorporeal membrane oxygenation (ECMO) has expanded for severe respiratory and circulatory failure. Acute kidney injury (AKI) is one of the most frequent complications. Up to 85% of ECMO patients develop AKI and approximately half of them require renal replacement therapy (RRT) making a comprehensive understanding of both therapies and their interaction essential for patient management. However, evidence to guide ECMO-specific RRT strategies remains limited. MAIN BODY: ECMO-related AKI arises from a complex interplay between patient and circuit factors that promote inflammation, endothelial injury, and tubular damage. Timing of RRT initiation in ECMO patients often relies on criteria used in the general critically ill population. Fluid overload is consistently associated with worse outcomes, and observational data suggest that earlier RRT, primarily to control fluid balance, may improve survival although randomized trials in ECMO patients are lacking. All RRT modalities can theoretically be used, but continuous techniques are preferred. RRT can be delivered via a parallel circuit using a dedicated venous catheter or integrated into the ECMO circuit. Integrated configurations reduce the need for additional vascular access and may prolong filter lifespan but require careful pressure management and team expertise to avoid alarms, hemolysis, and air embolism. Anticoagulation strategies must balance bleeding and thrombosis risk across both circuits; unfractionated heparin remains standard, while regional citrate anticoagulation can safely extend filter lifespan in selected patients although data is lacking in patients under veno-arterial ECMO. RRT during ECMO is associated with higher short-term mortality and an increased burden of chronic kidney disease among survivors. CONCLUSIONS: The management of AKI in ECMO patients remains a major clinical challenge. While RRT is often required in this population, optimal strategies for its initiation, modality selection, and integration with ECMO circuits are still evolving. Current evidence underscores the need for individualized approaches based on patient characteristics. Future research should focus on defining standardized protocols for RRT implementation in ECMO, use of regional citrate anticoagulation, optimizing patient selection, and evaluating long-term renal and survival outcomes.