Intensive Care Medicine Experimental最新文献

Background: Sepsis is characterized by organ dysfunction due to infection, with increasing evidence of mitochondrial dysfunction assessed preclinically and invasively. Protoporphyrin IX-triplet state lifetime technique (PpIX-TSLT) permits non-invasive determination of cellular oxygen metabolism and may provide deeper pathophysiological insights.

Methods: This analysis is part of a prospective monocentric cohort study. ICU patients with sepsis and septic shock and healthy controls were enrolled between May 2018 and June 2022. Mitochondrial oxygen tension (mitoPO₂), consumption (mitoVO₂) and delivery (mitoDO₂) were assessed in the skin of healthy controls and patients with sepsis in the acute phase (3 ± 1 days after onset) and long-term course of disease (6 ± 2 months after onset) using PpIX-TSLT (CE-certified Cellular Oxygen METabolism system). Primary endpoints were differences in mitoPO₂, mitoVO₂, and mitoDO₂ between patients in the acute phase of sepsis and controls. We tested group differences with t-tests and report Cohen's d (d) as effect size.

Results: In the acute phase, mitochondrial oxygen tension (mitoPO₂) was significantly reduced (n = 133, mean ± standard deviation: 58.4 ± 19.2 mmHg) compared to controls (n = 79, 67.3 ± 17.7 mmHg, p = 0.002, d = - 0.48). We found no significant differences in oxygen tension in the long-term course (n = 43) or in oxygen consumption and delivery between acute and long-term course of sepsis and controls. In the acute phase, lower mitochondrial oxygen delivery was associated with higher Sequential Organ Failure Assessment score (Spearman's ρ = - 0.23, p = 0.009) and higher lactate concentrations (ρ = - 0.21, p = 0.021) and, thus, correlated with disease severity.

Conclusions: Our results suggest that cellular oxygen metabolism in sepsis is characterized by a reversible restriction of oxygen tension without an impairment of mitochondrial oxygen consumption. Additionally, oxygen delivery is dependent on disease severity. These findings should be re-validated in a larger cohort.

Trial registration: NCT03620409 (Ethics vote: 5276-09/17; German Register of Clinical Studies: DRKS00013347), Principal investigator: Sina M. Coldewey, Date of Registration: 11-30-2017 NCT03620409.

Background: Delirium is a serious complication in patients with COVID-19-related acute respiratory distress syndrome (ARDS) admitted to the intensive care unit (ICU). Although numerous clinical risk factors have been identified, the immunologic pathways underlying delirium remain unclear. In this retrospective cohort study, we investigated high-dimensional immune signatures in ICU patients to delineate peripheral immune markers associated with delirium. We also explored machine learning (ML) approaches to enhance biomarker discovery and strengthen predictive modelling through synthetic data generation.

Methods: We studied a cohort of 62 COVID-19 ARDS patients admitted to the ICU at Lausanne University Hospital, Switzerland. The primary analysis compared patients within this cohort who developed delirium (n = 39) to those who remained delirium-free (n = 23). As a baseline for disease severity, we also compared the ICU cohort to 55 non-ICU COVID-19 patients and 450 healthy individuals. We performed high-dimensional immunophenotyping of cytokines, chemokines, and growth factors using multiplex beads assay, along with immune cell profiling via mass cytometry (CyTOF). Ridge regression has been employed to build classification models. We also generated synthetic samples using beta-variational autoencoders to improve sample size and subsequently model stability.

Results: Delirious patients exhibited a distinctive immune signature, including elevated CXCL1, CCL11, CXCL13, HGF, and VEGF-A, coupled with reduced IL-1α, IL-21, and IL-22. Alterations in immune cell populations featured increased exhausted B cells and decreases in CXCR3 + CD4 + T cells, IgM + unswitched memory B cells, and HLA-DR + activated T cells. Leveraging these high-dimensional data, we trained ridge regression models to predict delirium. Incorporating synthetic data helped stabilize the models with a best-performing model achieving an area under the curve (AUC) of 0.95, with high sensitivity (93%) and specificity (86%), based on 12 identified markers.

Conclusion: Our findings demonstrate a distinct immune profile linked to ICU delirium and illustrate how ML can enhance biomarker discovery. Further prospective validation may refine these markers and guide precision-targeted interventions for mitigating delirium in critically ill populations.

Background: Bronchopulmonary dysplasia-associated pulmonary hypertension (BPD-PH) seriously threatens the lives of preterm infants. The absence of animal models that can simulate its progression from early hyperoxic lung injury to late hypoxic vascular remodeling has hindered related research.

Objective: To establish a neonatal rat BPD-PH model by simulating exposure to sequential hyperoxic hypoxia experienced by human preterm infants.

Methods: Newborn SD rats were randomized into two control groups (C1 exposed to 21% O₂ for 2 weeks; C2 exposed to 21% O₂ for 3 weeks), and three exposure groups (H1 exposed to 75% O₂ for 2 weeks; H2 exposed to 75% O₂ for 2 weeks and then to 10% O₂ for a week; H3 exposed to 75% O₂ for 2 weeks and then to normoxia for a week). Cardiopulmonary parameters were evaluated by echocardiography, right ventricular systolic pressure measurement, histology, and α-SMA immunofluorescence.

Results: H1 and H2 groups exhibited distinct phenotypes, with those in the H2 group showing more severe phenotypes. The H2 group exhibited a 142% increase in RVSP relative to those in the C2 group. The right-heart index (RI) was 0.43 ± 0.01 in the H2 group, 0.36 ± 0.02 in the H3 group, and 0.22 ± 0.03 in the C2 group. Pulmonary vascular remodeling was significantly increased in the H2 group compared to the control and H3 groups. The H2 group uniquely replicated the disease process, with alveolar simplification preceding hypoxia-induced vascular thickening.

Conclusion: The sequential hyperoxic hypoxia model dynamically mimicked the clinical progression of BPD-PH, which may provide a powerful platform for stage-specific mechanism research and development of novel therapeutic strategies.

Background: Inotropic support is often used to improve hemodynamics and organ perfusion in patients with advanced heart failure-related cardiogenic shock (ADHF-CS). We aimed to evaluate the effect of inotrope timing on patient mortality in patients meeting Society for Cardiovascular Angiography and Interventions (SCAI) stage-C criteria within 24 h of hospital presentation.

Methods: We analyzed a local cardiogenic shock database of patients admitted to our cardiovascular intensive care units at the University of Pennsylvania from five emergency departments between 2021 and 2023. Adult patients with left ventricular ejection fraction ≤40% were eligible for inclusion. Patients with hypoperfusion, who met at least one physical, biochemical, and hemodynamic criterion for SCAI-C shock were included. The primary outcome was 28-day mortality. We also compared SCAI criteria and diagnostic examination timing between early and delayed inotropic support groups.

Results: A total of 138 out of 623 patients (22%) with cardiogenic shock met inclusion criteria for this study. 28-day mortality was higher in patients who received inotropic therapies ≥8 h after cardiogenic shock onset compared to patients who received earlier support (4-h odds ratio of death (OR) 3.19, 95% CI: 1.34-8.03; 8-h OR: 2.4, 95% CI: 1.09-5.26). 28-day mortality was lower in the early inotrope group (<8 h from shock onset) compared to the delayed (≥8 h) group (15/87; 17% vs. 17/51; 32%, p = 0.031). Patients with early inotropic support more often presented with a cool peripheral exam (34% vs. 16%, p = 0.022) and an initial lactate > 2 mmol/dL (71% vs. 49%, p = 0.009). Delayed inotropic support was associated with hypotension at presentation (84% vs. 57%, p = 0.001), longer time to echocardiography (19 [11-36] vs. 15 [3-24] h, p = 0.053) and time to pulmonary artery catheterization (25 [16-45] vs. 16 [2-46] h, p = 0.042).

Conclusion: Our findings suggest that inotropic support initiated within 8 h of acute presentation is associated with decreased 28-day mortality for patients with ADHF-related cardiogenic shock. Peripheral perfusion and cardiac output measurement were less frequently quantified within the first 24 h for patients with delayed inotropic support. Using shock classification tools, such as the SCAI shock criteria, may help identify patients with CS, especially in its early stages.

Background: Breath stacking, particularly double triggering, is a common patient-ventilator asynchrony during strong inspiratory effort. It can cause excessive tidal volumes and high transpulmonary pressures, contributing to ventilator-induced lung injury (VILI). The mode-specific consequences of breath stacking induced by strong inspiratory effort remain unclear.

Methods: In a porcine model of minimal lung injury, 17 animals were randomized to volume-controlled ventilation (VCV, n = 9) or pressure-controlled ventilation (PCV, n = 8). High respiratory drive was induced with continuous CO₂ inhalation, and ventilator settings were dynamically adjusted to maintain a breath stacking ratio of 40-70% of spontaneous efforts. Measurements included airway and transpulmonary pressures, driving pressures, tidal volume, esophageal pressure swings (ΔPes), stress index (SI), respiratory compliance, and histological lung injury. Risk factors for baro/volutrauma were defined by elevated plateau or driving pressures, transpulmonary pressures, or tidal volume >10 mL/kg. Atelectrauma risk was defined by SI < 0.9, negative end-expiratory transpulmonary pressure (PLexp), or vigorous effort (ΔPes > 5 cmH₂O or Pmus > 8 cmH₂O).

Results: VCV animals exhibited higher respiratory rates (44.0 vs. 30.5 breaths/min, p = 0.027), whereas PCV resulted in stronger inspiratory efforts (ΔPes 6.1 vs. 4.2 cmH₂O, p = 0.015). During breath stacking, VCV produced larger tidal volumes and higher plateau pressures, accumulating more baro/volutrauma risk factors (median 4.0 vs. 0.0, p < 0.001). In contrast, PCV animals developed more atelectrauma risk factors (3.0 vs. 1.0, p = 0.004). Histological injury scores were comparable, with a non-significant trend toward greater severity in PCV.

Conclusions: Breath stacking under strong inspiratory drive can promote lung injury through distinct mechanisms depending on ventilation mode. VCV was associated with the risk of overdistension, whereas PCV involved vigorous inspiratory effort and potential atelectrauma. Double triggering should be recognized as a clinical warning sign, prompting careful assessment of respiratory drive, inspiratory effort, and ventilator settings.

Background: Severe hypercapnia, especially when associated with acidosis, should be avoided or actively managed, as it is an independent risk factor in critically ill patients. The ADVOS multi hemodialysis system offers the potential to correct acidosis and hypercapnia through customizable pH and bicarbonate content within the dialysate fluid. The aim of this work is to analyze the exact timing for pH correction and the main factors leading to it in patients with multiple organ failure (MOF) and hypercapnic acidosis treated with ADVOS.

Methods: Patients with MOF and metabolic or hypercapnic acidosis (pH < 7.35) were included over a study period of 13 months. All patients received at least one treatment with ADVOS hemodialysis system for at least 24 h. The primary outcome was the time to blood pH ≥ 7.35.

Results: 24 patients with a median age of 61 years and a median SOFA score of 15 points were included. The results of 134 ADVOS sessions, with a median of 5 sessions per patient, were analyzed. Median time to blood pH ≥ 7.35 was 4 h; a significant blood pH increase within 24 h was reached in all patients (7.21 before vs. 7.39 after, p < 0.01).

Conclusions: A single session of ADVOS corrected blood pH and supported the reduction of pCO₂ with a median CO₂ removal of 55 mL/min in patients with multiple organ failure and hypercapnic acidosis.

Background: Transesophageal lung ultrasound (TELUS) has emerged as a novel modality that utilizes the esophageal acoustic window to obtain high-resolution images of posterior lung regions. However, its quantitative assessment and clinical relevance remain poorly explored. This study aimed to evaluate the feasibility and prognostic value of TELUS in critically ill patients, focusing on its association with arterial oxygenation and 28-day mortality.

Methods: In this prospective observational study, TELUS was performed in 69 mechanically ventilated ICU patients. TELUS imaging was acquired at three esophageal levels corresponding to posterior apical, mid, and basal lung regions. A semi-quantitative TELUS score was derived and its correlation with clinical variables was analyzed. Univariate and multivariate logistic regression analyses were employed to identify predictors of 28-day mortality. Receiver operating characteristic (ROC) analysis was used to assess the predictive performance of TELUS for mortality and severe hypoxemia (PaO₂/FiO₂ ≤ 100).

Results: Non-survivors had significantly higher TELUS scores compared to survivors (median 5 [IQR 4-6] vs. 4 [3-5], P = 0.001). Regional TELUS scores at the upper-aortic arch level and mid-esophageal level were elevated in non-survivors (P = 0.018 and P = 0.004, respectively). TELUS scores showed a significantly negative correlation with the PaO₂/FiO₂ ratio (r = -0.51, P < 0.0001), and positive correlations with PEEP (r = 0.32, P = 0.007) and SOFA scores (r = 0.26, P = 0.032). Multivariate analysis identified both SOFA (OR 1.31, 95% CI 1.08-1.63, P = 0.009) and TELUS scores (OR 1.72, 95% CI 1.08-2.96, P = 0.030) as independent predictors of 28-day mortality. ROC analysis showed that a TELUS score ≥ 4 predicted 28-day mortality and severe hypoxemia (PaO₂/FiO₂ ≤ 100), yielding areas under the ROC (AUCs) of 0.74 and 0.72, with high sensitivity (89% and 100%, respectively) and negative predictive values (92% and 100%, respectively.) CONCLUSION: TELUS is a feasible novel technique that provides a reliable assessment of posterior lung aeration in critically ill patients. TELUS scoring correlates with impaired oxygenation and is independently associated with 28-day mortality. These findings highlight the prognostic value of TELUS and support its potential integration into transesophageal cardiopulmonary ultrasound protocols.

Background: Early identification of acute kidney injury (AKI) in the intensive care unit (ICU) remains challenging. We aimed to identify key predictors of new-onset AKI within 48 h after ICU admission and renal replacement therapy (RRT) need within 7 days, using explainable artificial intelligence (XAI) with eXtreme Gradient Boosting (XGBoost). We also assessed whether XGBoost improved predictive performance.

Methods: A retrospective cohort study across four ICUs was conducted as part of the SWECRIT biobank project. Blood samples were prospectively obtained at ICU admission and retrospectively analysed. AKI was defined by the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. XAI models were compared with logistic regression models, incorporating emerging biomarkers and routine clinical data at ICU admission. SHapley Additive exPlanations (SHAP) were used to identify key predictors. Discrimination was assessed using the mean area under the receiver operating characteristic curve (AUC).

Results: The study included 4732 admissions, with 2603 analysed for new-onset AKI and 4716 for RRT. Top predictors of new-onset AKI were urine output, endostatin, baseline creatinine, lactate, and albumin. Top predictors of RRT were creatinine, urine output, endostatin, neutrophil gelatinase-associated lipocalin (NGAL), and the Simplified Acute Physiology Score (SAPS) 3. Several clinically relevant non-linear relationships were revealed. XGBoost outperformed logistic regression for both new-onset AKI (mean AUC 0.76, 95% CI 0.70-0.81 vs. 0.74, 95% CI 0.68-0.81; p < 0.001) and RRT (mean AUC 0.92, 95% CI 0.89-0.95 vs. 0.90, 95% CI 0.87-0.94; p < 0.001).

Conclusion: XGBoost identified key predictors of early new-onset AKI and RRT need in the ICU, highlighting both emerging (endostatin, NGAL) and established biomarkers (lactate, albumin), alongside known clinical predictors. It also improved predictive accuracy for both outcomes. Further clinical evaluation of these biomarkers and XAI models is warranted.

Background: Induced hypothermia after cardiac arrest is neuroprotective in several animal models of cardiac arrest, but few high-quality studies have been conducted in larger animals. Recent clinical trials have questioned the beneficial effects of post-ischemic hypothermia. This study investigated whether immediate cooling or a 2-h delay in cooling to 33 °C after cardiac arrest was neuroprotective compared to controlled normothermia in large animals.

Methods: Young adult female swine were anesthetized and kept at normothermia (38 °C). All animals were subject to 10 min of cardiac arrest by ventricular fibrillation, followed by 4 min of cardiopulmonary resuscitation, before the first countershock. At 10 min of return of spontaneous circulation (ROSC), animals were included and randomized to receive immediate hypothermia (33 °C), 2-h delayed hypothermia (33 °C), or normothermia for 30 h, including both cooling and rewarming time. Animals were extubated and assessed for 7 days. The primary outcome was brain histopathology using a modified Histology Damage Score. Secondary outcomes were neurocognitive testing, neurologic deficit score, and biomarkers of brain injury.

Results: Among 42 animals, 33 were included; 11 in each arm, 23 survived until day 7. The modified Histology Damage Score was not significantly different between groups (p = 0.29). Neither neurocognitive testing nor neurologic deficit scores showed significant differences between the groups (p = 0.11 and p = 0.67, respectively). Neurofilament light chain (NfL) levels were significantly lower in the immediate hypothermia group at 48 h and on day 7 compared to the normothermia group (p = 0.0087, p = 0.012), but not in the delayed hypothermia group (p = 0.075, p = 0.33).

Conclusion: Our experimental model in large swine showed no neuropathological or functional protective effect of induced hypothermia after cardiac arrest, but NfL levels were lower in animals receiving immediately induced hypothermia, suggesting mitigation of neuronal injury.

Trial registry: Preclinicaltrials.eu (PCTE0000272), published 2021-11-03.