Critical Care Medicine最新文献

Objectives: Real-time automated alert and trigger systems that notify clinical staff beyond the patient's assigned bedside nurse are increasingly used to improve the recognition and response to clinical deterioration in hospital wards. This review evaluated their effects on clinical outcomes in adult general ward patients.

Data sources: Seven electronic databases were searched from inception to April 1, 2024, supplemented with citation tracking.

Study selection: Two sets of reviewers independently screened for comparative studies assessing automated clinical deterioration alert and trigger systems vs. conventional manual escalation processes.

Data extraction: Two sets of reviewers extracted data. Primary outcomes were hospital mortality and in-hospital cardiac arrest. Secondary outcomes included unplanned ICU transfers, and hospital and ICU length of stay. Data were pooled using random-effects meta-analysis.

Data synthesis: Eighteen studies ( n = 349,818) were included: two randomized controlled trials (RCTs), 15 before-and-after designs, and one quasi-experimental study. All studies had at least moderate risk of bias. The main meta-analysis included nine studies ( n = 58,632) involving alert-eligible cohorts (i.e., patients meeting alert thresholds). Based on two before-and-after studies, automated alert and trigger systems were associated with a significant reduction in-hospital cardiac arrests (risk ratio [RR], 0.60; 95% CI, 0.43-0.85; p = 0.004). No significant reduction in hospital mortality was observed (RR, 0.80; 95% CI, 0.62-1.05; p = 0.09), based on pooled data from RCTs and before-and-after studies. ICU length of stay was also significantly reduced, based on one RCT and one before-and-after study. No significant effects were found for hospital length of stay or unplanned ICU transfers in either alert-eligible or all-admission cohorts.

Conclusions: Real-time clinical deterioration automated alert and trigger systems are associated with reduced in-hospital cardiac arrests and may offer mortality benefits when integrated into structured escalation pathways. Further high-quality trials are needed to confirm these findings and optimize system design.

Objectives: We sought to determine trends in use of IV vitamin C for hospitalized patients with sepsis in the context of evolving evidence, including a single-center before-after study in late 2016 and several trials in 2019-2021.

Design: Retrospective cohort study.

Setting: One thousand one hundred fifteen U.S. hospitals contributing to the Premier Healthcare Database, 2008-2021.

Patients: Eleven million three hundred seventy-five thousand three hundred twenty-six adult inpatients with sepsis.

Interventions: IV vitamin C, at any point of the hospital stay.

Measurements and main results: Patients had a median (interquartile range [IQR]) age of 71 years (59-81 yr) and a median (IQR) of 5 comorbidities (4-7 comorbidities); 53.0% were female; on hospital day 1, 6.9% were mechanically ventilated and 7.5% received a vasopressor. Overall, 32,131 patients (0.3%) received IV vitamin C at any point during hospitalization. During the study period, administration fell from 2008, quarter 1 (0.5%) through 2017, quarter 1 (< 0.1%), then rose and peaked in 2020, quarter 1 (0.6%), and fell through 2021, quarter 4 (0.1%). Examining three time periods defined by predetermined cutpoints (2015 quarter 4, when International Classification of Diseases coding for sepsis changed, and 2020 quarter 1, when the COVID-19 pandemic began), vitamin C use also varied ( p < 0.001): 0.2% (2008 quarter 1 to 2015 quarter 3); 0.3% (2015 quarter 4 to 2019 quarter 4); and 0.3% (2020-2021). Temporal trends were similar in sicker subcohorts defined by early mechanical ventilation, early vasopressor use, and diagnosis of COVID-19 (2020-2021). A multilevel logistic regression model with data from 91 hospitals that contributed at least 1 sepsis case per quarter showed a similar utilization pattern, with substantial between-hospital variability (median odds ratio, 7.78; 95% CI, 5.45-11.58).

Conclusions: IV vitamin C prescription for hospitalized patients with sepsis in the United States was overall infrequent over the 14-year study period, rising after the publication of a before-after study and declining in the COVID-19 pandemic as clinical trial results emerged.

Objectives: High-flow nasal cannula is increasingly used in patients with acute hypoxemic respiratory failure. However, its impact on diaphragmatic function remains poorly understood.

Design: Observational crossover pilot study.

Setting: Single-center ICU.

Patients: Twenty-eight adult ICU patients with a Pa o2 /F io2 ratio less than or equal to 300 mm Hg were enrolled.

Interventions: Patients underwent sequential 60-minute phases of noninvasive respiratory support using a Venturi mask, high-flow nasal cannula (40 L/min), and helmet continuous positive airway pressure (CPAP).

Measurements and main results: Diaphragmatic function was assessed using ultrasound, while inspiratory effort was evaluated through esophageal pressure swings. Arterial blood gases were also collected. High-flow nasal cannula significantly improved the diaphragmatic thickening fraction compared with the Venturi mask (27% ± 9.9% vs. 20% ± 6%; p = 0.0013). Conversely, diaphragmatic excursion was lower with high-flow nasal cannula than with both the Venturi mask and CPAP (1.1 ± 0.63 cm vs. 1.5 ± 0.95 cm and 1.4 ± 0.59 cm, respectively; p = 0.0002). High-flow nasal cannula also reduced inspiratory effort compared with the Venturi mask. In patients with diaphragmatic dysfunction index greater than 100, both high-flow nasal cannula and CPAP enhanced diaphragmatic thickening and decreased esophageal pressure swings relative to the Venturi mask.

Conclusions: This study shows that high-flow nasal cannula improves diaphragmatic function compared with Venturi mask oxygen therapy. Larger studies are needed to confirm these findings.

Objectives: Trauma-induced coagulopathy biomarkers may be influenced by injury mechanism. We sought to identify differences in patterns of coagulopathy with and without severe traumatic brain injury (TBI).

Design: Retrospective cohort study.

Setting: Harmonized database composed of six major hemorrhagic shock trials: Control of Major Bleeding After Trauma (COMBAT), Cold-stored Platelet Early Intervention in Hemorrhagic Shock (CriSP-HS), Prehospital Air Medical Plasma (PAMPer), Prehospital Whole Blood in Emergency Resuscitation (PPOWER), Pragmatic, Randomized Optimal Platelet and Plasma Ratios (PROPPR), and Study of Tranexamic Acid During Air Medical and Ground Prehospital Transport (STAAMP).

Patients: All subjects randomized to placebo or standard-of-care groups with complete data for international normalized ratio (INR), thromboelastography values (alpha angle [AA], K time, maximum amplitude [MA], and lysis in 30 min), and Abbreviated Injury Scores (AISs). Subjects from COMBAT and CriSP-HS were screened and ultimately excluded from the final analysis as they did not meet eligibility criteria.

Interventions: None.

Measurements and main results: Stratified principal component (PC) analysis was performed for INR and thromboelastography values. Strata were defined based on AIS scores as: 1) isolated severe TBI (iTBI); 2) severe polytrauma (POLY), those with both major head injury and torso/extremity trauma; and 3) isolated severe torso/extremity trauma (iTRUNK). We identified 506 subjects with complete data: 51 iTBI, 115 POLY, and 340 iTRUNK. For each stratum, two PCs were identified that accounted for more than 65% of the variance. Associations between PC scores and outcomes, including need for blood product transfusion within 24 hours as a surrogate marker for early coagulopathy and bleeding, were examined with logistic regression. For both iTBI and POLY, PC1 included INR, AA, K time, and MA, and was associated with greater odds of early transfusion (odds ratio [OR], 3.57; 95% CI, 1.50-8.45; p = 0.004 for iTBI and OR, 2.29; 95% CI, 1.11-4.75; p = 0.026 for POLY). For iTRUNK, PC1 included INR, AA, and MA and was protective with reduced odds of early transfusion (OR, 0.51; 95% CI, 0.37-0.70; p < 0.001).

Conclusions: PC analysis demonstrated a unique pattern of coagulation biomarkers common to patients with severe TBI, irrespective of other injuries.

Rationale: Neuromuscular blocking agents (NMBAs) show potential benefits on mortality and other complications of acute respiratory distress syndrome (ARDS) in adult patients. Evidence-based decisions and processes ensure appropriate use of neuromuscular blockade in adult patients with ARDS.

Objectives: The objective of these guidelines was to develop evidence-based recommendations for the administration of NMBAs in critically ill adult patients with ARDS.

Design: The American College of Critical Care Medicine Board convened a 21-member multidisciplinary panel of experts in critical care medicine, nursing, respiratory therapy, pharmacology, surgery, neurology, and anesthesiology. The panel included two expert methodologists specialized in developing evidence-based recommendations in alignment with the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) methodology. Conflict-of-interest policies were strictly followed during all phases of guidelines development including task force selection and voting.

Methods: The panel members identified and formulated five Population, Intervention, Comparison, and Outcome questions. We conducted a systematic review for each question to identify the best available evidence, statistically analyzed the evidence, and assessed the certainty of the evidence using the GRADE methodology. We used the GRADE evidence-to-decision framework to formulate the recommendations.

Results: The panel generated two conditional recommendations. One recommendation is to use NMBAs in adults with ARDS with Pao2/Fio2 less than 150. For the other recommendations, there was equipoise in the recommendation for and against using titratable vs. fixed-dose NMBA dosing, a monitoring-based strategy for assessing depth of sedation and analgesia in adults with ARDS before initiating or while receiving neuromuscular blockade, and administration of NMBAs for patients who are proned, due to overall lack of evidence in critically ill patients and due to considerations of patient safety and experience concerns.

Conclusions: These guidelines provide additional perspectives on the use of NMBA in patients with ARDS, recognizing that institutional and patient-specific considerations must help to guide the decision-making process.

Objectives: Acute kidney injury (AKI) and delirium are common complications of critical illness. However, relatively few studies have evaluated their relationship. We conducted a systematic synthesis and meta-analysis of existing evidence to clarify this association in critically ill patients.

Data sources: A comprehensive search was conducted across MEDLINE, Embase, CINAHL, Scopus, Web of Science, and Cochrane Library for publications reporting both AKI and delirium in ICUs patients from January 2000 to January 2025.

Study selection: AKI was defined according to serum creatinine or urine output criteria based on the contemporary definitions used in the individual studies. The primary outcome was the proportion of critically ill patients with AKI who developed delirium. Secondary outcomes included mortality and health service utilization.

Data extraction: Pooled meta-analyses were summarized as effect sizes in proportions, risk ratios (RRs), odds ratios, or weighted mean differences (WMDs) using a random-effects model. The certainty of evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation approach.

Data synthesis: Eighteen observational studies comprising 158,694 patients were included. Overall study quality was moderate. The pooled proportion of delirium among patients with AKI was 32% (95% CI, 18-47%). Delirium was associated with higher mortality (RR, 2.36; 95% CI, 1.61-3.47; moderate certainty), greater renal replacement therapy use (RR, 3.12; 95% CI, 1.89-5.15; moderate certainty), longer ICU stays (WMD, 3.54 d; 95% CI, 1.20-5.87 d; moderate certainty), and longer hospital stays (WMD, 4.78 d; 95% CI, 3.48-6.09 d; moderate certainty) compared with patients with AKI not experiencing delirium.

Conclusions: Delirium is common among critically ill patients with AKI and is associated with worse outcomes and greater health resource use.

Objectives: To examine how accurately ICU clinicians perceived family-reported prognostic expectations (FPEs) for patients with prolonged mechanical ventilation (PMV).

Design: A cross-sectional, exploratory design using secondary analysis.

Setting: Thirteen ICUs across five hospitals in the United States.

Subjects: Family members of patients with PMV and ICU clinicians, including physicians and nurses.

Interventions: None.

Measurements and main results: Latent profile analysis was used to identify profiles of accuracy in clinician perception of FPE, followed by bivariate analyses and multinomial logistic regression to examine associations between patient, family, and clinician characteristics and profile membership. A total of 554 participants (239 family members, 150 physicians, and 165 nurses) were included. Five distinct latent profiles of accuracy in clinician perception of FPE were identified: 1) clinician underestimation of FPE; 2) clinician overestimation of FPE; 3) accurate perception: low prognosis; 4) accurate perception: moderate prognosis; and 5) accurate perception: high prognosis. Families in profile 1 (clinician underestimation of FPE) were more likely to be spouses/partners of patients and reported higher levels of hope and optimism, whereas those in profile 2 (clinician overestimation of FPE) reported lower levels. Patient characteristics, including age, employment status, admission to medical ICU, and pulmonary-related hospital diagnosis, were statistically significantly associated with the profile membership.

Conclusions: Understanding how accurately clinicians perceive FPE is vital to improving shared decision-making and developing goal-concordant care for patients with PMV. Further research examining strategies for clinicians to accurately perceive what families believe about prognosis is needed to identify potential misalignment, initiate timely and empathetic conversations, and build toward shared decision-making and goal-concordant care.

Objectives: To evaluate evolving management, ICU admission, and outcomes for critically ill chimeric antigen receptor (CAR) T-cell patients over a 6-year period.

Design: Multicenter retrospective cohort study from January 2018 to September 2023.

Setting: Eight U.S. centers.

Patients: Adult CAR T-cell patients requiring ICU admission.

Interventions: None.

Methods: Summary statistics included mean, sd , median, and interquartile range (IQR). Fisher exact test or chi-square test were used to evaluate association between year treated and other categorical variables. Cochran-Armitage test was performed to assess significance of trends across years. Multivariable logistic regression was performed to assess covariates associated with mortality.

Measurements and main results: Demographics, toxicity management, ICU admission, support modalities, toxicity severity, and survival (ICU, hospital, and 3-mo) were compared year-to-year. From 2018 to 2023, 2238 patients received CAR T cells, with increasing number of patients treated yearly; 398 (17.8%) required ICU care. Of those admitted to the ICU, 66.1% were male, 89.2% had lymphoma, and median age was 64 years (53-71 yr). ICU admission rates declined from 38.5% (95% CI, 31.6-45.8%) in 2018 to 16.4% in 2023 (95% CI, 13.5-19.7%; p < 0.0001). Cytokine release syndrome or immune effector cell-associated neurotoxicity syndrome was the reason for ICU admission in 87.9%. In 2023 vs. 2018, ICU patients were older (median, 65 yr [IQR, 55-73 yr] vs. 58 yr [48-67 yr]; p = 0.003) with higher comorbidity indices (4 [4-6] vs. 3 [2-4]; p = 0.005) and more severe toxicities (≥ grade 3: 90.1% vs. 69.9%; p = 0.004). Corticosteroid use for less severe toxicities (≤ grade 2 toxicity: 73.8% vs. 40.6%; p = 0.0001) and anakinra use (56% vs. 5.5%; p < 0.0001) increased throughout the years. Mortality from cytokine release syndrome and immune effector cell-associated neurotoxicity syndrome remained low (5.5%). Age, Sequential Organ Failure Assessment greater than or equal to 10 at ICU admission, and ICU admission for noncytokine release/neurotoxicity syndrome reasons were associated with hospital mortality (odds ratios, 1.02 [95% CI, 1-1.04; p = 0.046], 4.69 [2.44-9.01; p < 0.0001], and 3.74 [1.91-7.3; p = 0.0001], respectively).

Conclusions: ICU admission rates after CAR T-cell treatment are declining. Although ICU patients are older with higher severity of illness and toxicity grades, ICU mortality after CAR T-cell therapy remains low.