Impact of Critical Illness on Resource Utilization: A Comparison of Use in the Year Before and After ICU Admission* Objectives: Increasingly, patients admitted to an ICU survive to hospital discharge; many with ongoing medical needs. The full impact of an ICU admission on an individual’s resource utilization and survivorship trajectory in the United States is not clear. We sought to compare healthcare utilization among ICU survivors in each year surrounding an ICU admission. Design: Retrospective cohort of patients admitted to an ICU during one calendar year (2012) in a multipayer healthcare system. We assessed mortality, hospital readmissions (categorized by ambulatory care sensitive conditions and emergency department), and outpatient visits. We compared the proportion of patients with visits during the pre-ICU year versus the post-ICU year. Patients: People admitted to an Intermountain healthcare ICU for greater than 48 hours in the year 2012 Interventions: None. Measurements and Main Results: Among 4,074 ICU survivors, 45% had increased resource utilization. Readmission rates at 30-day, 90-day, and 1-year were 15%, 26%, and 43%. The proportion of patients with a hospital admission increased significantly in the post-ICU period (43% vs 29%; p < 0.001). Of patients with a readmission in the post-ICU period, 24% were ambulatory care sensitive condition. Patients with increased utilization differed by socioeconomic status, insurance type, and severity of illness. Sixteen percent of patients had either an emergency department or inpatient admission, but no outpatient visits during the post-ICU period. Conclusions: An ICU admission is associated with increased resource utilization including hospital readmissions, with many due to an ambulatory care sensitive condition. Lower socioeconomic status and higher severity of illness are associated with increased resource utilization. After an ICU visit patients seem to use hospital resources over outpatient resources. Interventions to improve and coordinate care after ICU discharge are needed. Drs. Hirshberg, Hopkins, and Brown and Ms. Wilson conceptualize the article. Drs. Hirshberg, Kuttler, and Brown, Ms. Wilson, and Ms. Stanfield acquire the data. Drs. Hirshberg, Majercik, Beesley, Orme, and Brown and Ms. Wilson analyze the data. Drs. Hirshberg, Majercik, Beesley, Orme, Hopkins, and Brown write the article. All authors approve the final copy and revise the article for important intellectual content. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Supported, in part, by grant from the departmental funds. Dr. Hopkins’ institution received funding from Intermountain Research and Medical Foundation to study outcomes of family caregivers of ICU patients. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: Ellie.hirshberg@imail.org; ellie.hirshberg@hsc.utah.edu This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Clinical Practice Guidelines and Consensus Statements About Pain Management in Critically Ill End-of-Life Patients: A Systematic Review Objectives: To identify and synthesize available recommendations from scientific societies and experts on pain management at the end-of-life in the ICU. Data Sources: We conducted a systematic review of PubMed, EMBASE, the Cochrane Database of Systematic Reviews, and Biblioteca Virtual en Salud from their inception until March 28, 2019. Study Selection: We included all clinical practice guidelines, consensus statements, and benchmarks for quality. Data Extraction: Study selection, methodological quality, and data extraction were performed independently by two investigators. A quality assessment was performed by four investigators using the Appraisal of Guidelines for Research and Evaluation II instrument. The recommendations were then synthesized and categorized. Data Synthesis: Ten publications were included. The Appraisal of Guidelines for Research and Evaluation II statement showed low scores in various quality domains, especially in the applicability and rigor of development. Most documents were in agreement on five topics: 1) using a quantitative tool for pain assessment; 2) administering narcotics for pain relief and benzodiazepines for anxiety relief; 3) against prescribing neuromuscular blockers during withdrawal of life support to assess pain; 4) endorsing the use of high doses of opioids and sedatives for pain control, regardless of the risk that they will hasten death; and 5) using quality indicators to improve pain management during end-of-life in the ICU. Conclusions: In spite of the lack of high-quality evidence, recommendations for pain management at the end-of-life in the ICU are homogeneous and are justified by ethical principles and agreement among experts. Considering the growing demand for the involvement of palliative care teams in the management of the dying patients in the ICU, there is a need to clearly define their early involvement and to further develop comprehensive evidence-based pain management strategies. Based on the study findings, we propose a management algorithm to improve the overall care of dying critically ill patients. Drs. Durán-Crane and Laserna contributed equally. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Dr. Durán-Crane disclosed he was previously employed by Fundacion Santa Fe de Bogotá, Colombia. Dr. Lopez-Olivo’s work is supported by a career award from the Rheumatology Research Foundation. Dr. Nates’ institution received funding from the George Sweeney Fellowship grant of the MD Anderson resources grant. The remaining authors have disclosed that they do not have any potential conflicts of interest. Systematic Review Registration: PROSPERO Number CRD42018089489. For information regarding this article, E-mail: jlnates@mdanderson.org Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Circuit Hemodynamics and Circuit Failure During Continuous Renal Replacement Therapy Objectives: To study hemodynamic changes within continuous renal replacement therapy circuits and evaluate their relationship with continuous renal replacement therapy longevity. Design: Analysis of downloaded variables recorded by continuous renal replacement therapy machines during multiple episodes of clinical care. Setting: Tertiary ICU in Melbourne, Australia. Patients: Cohort of 149 ICU patients: 428 episodes of continuous renal replacement therapy. Interventions: None. Measurements and Main Results: Indices of continuous renal replacement therapy function representing 554,991 minutes were assessed including blood flow, access pressure, effluent pressure, prefilter pressure, and return pressure. We defined three patterns of artificial kidney failure: early (≤ 12 hr), intermediate (> 12–24 hr), and late (> 24 hr) in 35%, 31%, and 34% of circuits, respectively. Mean access pressure in late artificial kidney failure was 7.5 mm Hg (7.1–7.9 mm Hg) less negative than early failing circuits and pressures demonstrated lower variability in such late failing circuits. Access dysfunction, defined as access pressure less than or equal to –200 mm Hg occurred in the first 4 hours in 118 circuits (27%) which had a shorter (median [interquartile range]) life at 12.9 hr [5.5–21.3 hr]) hours than access dysfunction-free circuits (18.8 hr [10.1–33.4 hr]; p < 0.0001). Multivariate analysis found the first occurrence of access dysfunction (as a time-varying covariate) was independently associated with increased hazard of subsequent failure (hazard ratio, 1.75; 1.36–2.26). Classification and regression tree analysis of summary pressure indices in the first 2 hours confirmed minimum access pressure to be a significant predictor, as well as indices of transmembrane pressure and return pressure. A pressure-based predictor correctly identified early and late failing circuits (86.2% and 93.6% specificity, respectively). Conclusions: Access dysfunction is a predictor of continuous renal replacement therapy circuit failure. Future monitoring of continuous renal replacement therapy hemodynamics may facilitate remedial actions to improve circuit function. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Dr. Sansom receives a research scholarship from the University of Melbourne. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: rinaldo.bellomo@mh.org.au Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Airway Pressure Release Ventilation in Adult Patients With Acute Hypoxemic Respiratory Failure: A Systematic Review and Meta-Analysis Objectives: To evaluate the efficacy and safety of airway pressure release ventilation in critically ill adults with acute hypoxemic respiratory failure. Data Sources: A systematic literature search of MEDLINE via PUBMED, EMBASE, the Cochrane Library, published conference proceedings and abstracts, reference lists of eligible studies and review articles, and hand searches of relevant journals and trial registers. Study Selection: Eligible studies included randomized controlled trials published between years 2000 and 2018, comparing airway pressure release ventilation to any ventilation mode, in critically ill adults with acute hypoxemic respiratory failure and reporting at least one mortality outcome. Data Extraction: Screened citations were reviewed and extracted independently by two investigators onto a prespecified proforma. Data Synthesis: There were 412 patients from seven randomized controlled trials included in the qualitative and quantitative data synthesis. Airway pressure release ventilation was associated with a significant mortality benefit (relative risk, 0.67; 95% CI, 0.48–0.94; I2 < 0.1%; p = 0.97) and improvement in day 3 PaO2/FIO2 ratio (weighted mean difference, 60.4; 95% CI, 10.3–110.5). There was no significant difference in requirement to initiate rescue treatments including inhaled pulmonary vasodilators, prone positioning, or extracorporeal membrane oxygenation (relative risk, 0.51; 95% CI, 0.22–1.21; I2 = 64.7%; p = 0.04). The risk of barotrauma was only reported in three studies and did not differ between groups (relative risk, 0.39; 95% CI, 0.12–1.19; I2 < 0.1%; p = 0.99). Conclusions: In adult patients requiring mechanical ventilation for acute hypoxic respiratory failure, airway pressure release ventilation is associated with a mortality benefit and improved oxygenation when compared with conventional ventilation strategies. Given the limited number of patients enrolled in the available studies, larger multicenter studies are required to validate these findings. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: ed_litton@hotmail.com Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Feasibility of Hidden Markov Models for the Description of Time-Varying Physiologic State After Severe Traumatic Brain Injury Objectives: Continuous assessment of physiology after traumatic brain injury is essential to prevent secondary brain insults. The present work aims at the development of a method for detecting physiologic states associated with the outcome from time-series physiologic measurements using a hidden Markov model. Design: Unsupervised clustering of hourly values of intracranial pressure/cerebral perfusion pressure, the compensatory reserve index, and autoregulation status was attempted using a hidden Markov model. A ternary state variable was learned to classify the patient’s physiologic state at any point in time into three categories (“good,” “intermediate,” or “poor”) and determined the physiologic parameters associated with each state. Setting: The proposed hidden Markov model was trained and applied on a large dataset (28,939 hr of data) using a stratified 20-fold cross-validation. Patients: The data were collected from 379 traumatic brain injury patients admitted to Addenbrooke’s Hospital, Cambridge between 2002 and 2016. Interventions: Retrospective observational analysis. Measurements and Main Results: Unsupervised training of the hidden Markov model yielded states characterized by intracranial pressure, cerebral perfusion pressure, compensatory reserve index, and autoregulation status that were physiologically plausible. The resulting classifier retained a dose-dependent prognostic ability. Dynamic analysis suggested that the hidden Markov model was stable over short periods of time consistent with typical timescales for traumatic brain injury pathogenesis. Conclusions: To our knowledge, this is the first application of unsupervised learning to multidimensional time-series traumatic brain injury physiology. We demonstrated that clustering using a hidden Markov model can reduce a complex set of physiologic variables to a simple sequence of clinically plausible time-sensitive physiologic states while retaining prognostic information in a dose-dependent manner. Such states may provide a more natural and parsimonious basis for triggering intervention decisions. Drs. Smielewski and Ercole are joint senior authors. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The software for brain monitoring ICM+ is licensed by the University of Cambridge (Cambridge Enterprise). Drs. Czosnyka and Smielewski received funding from Cambridge Enterprise (part of license fees for the brain monitoring software ICM+ used in this project for data collection and preprocessing); also supported by National Institute for Health Research Cambridge Brain Repair Centre. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: Shadnaz.Asgari@csulb.edu Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Determining the Association Between End-of-Life Care Resources and Patient Outcomes in Pennsylvania ICUs Objectives: As ICUs are increasingly a site of end-of-life care, many have adopted end-of-life care resources. We sought to determine the association of such resources with outcomes of ICU patients. Design: Retrospective cohort study. Setting: Pennsylvania ICUs. Patients: Medicare fee-for-service beneficiaries. Interventions: Availability of any of one hospital-based resource ( palliative care consultants) or four ICU-based resources ( protocol for withdrawal of life-sustaining therapy, triggers for automated palliative care consultation, protocol for family meetings, and palliative care clinicians embedded in ICU rounds). Measurements and Main Results: In mixed-effects regression analyses, admission to a hospital with end-of-life resources was not associated with mortality, length of stay, or treatment intensity (mechanical ventilation, hemodialysis, tracheostomy, gastrostomy, artificial nutrition, or cardiopulmonary resuscitation); however, it was associated with a higher likelihood of discharge to hospice (odds ratio, 1.58; 95% CI, 1.11–2.24), an effect that was driven by ICU-based resources (odds ratio, 1.37; 95% CI, 1.04–1.81) rather than hospital-based resources (odds ratio, 1.19; 95% CI, 0.83–1.71). Instrumental variable analysis using differential distance (defined as the additional travel distance beyond the hospital closest to a patient’s home needed to reach a hospital with end-of-life resources) demonstrated that among those for whom differential distance would influence receipt of end-of-life resources, admission to a hospital with such resources was not associated with any outcome. Conclusions: ICU-based end-of-life care resources do not appear to change mortality but are associated with increased hospice utilization. Given that this finding was not confirmed by the instrumental variable analysis, future studies should attempt to verify this finding, and identify specific resources or processes of care that impact the care of ICU patients at the end of life. This work was performed at the Perelman School of Medicine at the University of Pennsylvania in Philadelphia, Pennsylvania. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Dr. Ashana’s institution received funding from the National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI) T32 HL007891. Drs. Ashana, Kohn, Madden, Harhay, and Kerlin received support for article research from the NIH. Dr. Umscheid’s institution received funding from Agency for Healthcare Research and Quality Evidence-based Practice Center contracts and the U.S. Food and Drug Administration, and he received funding from the Patient-Centered Outcomes Research Institute Advisory Panel. Dr. Stephens-Shields received funding from Juniper Pharmaceuticals, October 2016 to March 2017. Dr. Harhay’s institution received funding from the NIH/NHLBI K99 HL141678. Dr. Kerlin’s institution received funding from the NIH/NHLBI K08 HL116771. Dr. Chen disclosed that he does not have any potential conflicts of interest. For information regarding this article, E-mail: deepshikha.ashana@uphs.upenn.edu Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Epidemiology and Outcome of Critically Ill Pediatric Cancer and Hematopoietic Stem Cell Transplant Patients Requiring Continuous Renal Replacement Therapy: A Retrospective Nationwide Cohort Study Objective: Acute kidney injury requiring continuous renal replacement therapy is a serious treatment-related complication in pediatric cancer and hematopoietic stem cell transplant patients. The purpose of this study was to assess epidemiology and outcome of these patients requiring continuous renal replacement therapy in the PICU. Design: A nationwide, multicenter, retrospective, observational study. Setting: Eight PICUs of a tertiary care hospitals in the Netherlands. Patients: Pediatric cancer and hematopoietic stem cell transplant patients (cancer and noncancer) who received continuous renal replacement therapy from January 2006 to July 2017 in the Netherlands. Interventions: None. Measurement and Main Results: Of 1,927 PICU admissions of pediatric cancer and hematopoietic stem cell transplant patients, 68 of 70 evaluable patients who received continuous renal replacement therapy were included. Raw PICU mortality was 11.2% (216/1,972 admissions). PICU mortality of patients requiring continuous renal replacement therapy was 54.4% (37/68 patients). Fluid overload (odds ratio, 1.08; 95% CI, 1.01–1.17) and need for inotropic support (odds ratio, 6.53; 95% CI, 1.86–23.08) at the start of continuous renal replacement therapy were associated with PICU mortality. Serum creatinine levels increased above 150% of baseline 3 days before the start of continuous renal replacement therapy. Urine production did not reach the critical limit of oliguria. In contrast, body weight (fluid overload) increased already 5 days prior to continuous renal replacement therapy initiation. Conclusions: PICU mortality of pediatric cancer and hematopoietic stem cell transplant patients requiring continuous renal replacement therapy is sadly high. Fluid overload at the initiation of continuous renal replacement therapy is the most important and earliest predictor of PICU mortality. Our results suggest that the most commonly used criteria of acute kidney injury, that is, serum creatinine and urine production, are not useful as a trigger to initiate continuous renal replacement therapy. This highlights the urgent need for prospective studies to generate recommendations for effective therapeutic interventions at an early phase in this specific patient population. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: p.a.m.a.raymakers-janssen@umcutrecht.nl This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Nonpharmacologic Interventions to Prevent or Mitigate Adverse Long-Term Outcomes Among ICU Survivors: A Systematic Review and Meta-analysis Objective: ICU survivors suffer from long-lasting physical, mental, and cognitive health impairments, also called “postintensive care syndrome”. However, an overview of the effectiveness of interventions to prevent or mitigate these impairments is lacking. The aim of this study is to assess the effectiveness of nonpharmacologic interventions. Data Sources: PubMed, CINAHL, PsycINFO, Embase, and Cochrane Library were systematically searched from inception until July 19, 2018. Study Selection: (Non)randomized clinical trials, controlled before-after studies, and interrupted time series were included. Outcomes of interest included patients physical, mental and cognitive outcomes, quality of life, and outcomes such as social functioning and functional status, measured after hospital discharge. Data Extraction: Two independent reviewers selected studies, extracted data, and assessed the risk of bias. Pooled mean differences and standardized mean differences were calculated using random-effect meta-analyses. Data Synthesis: After screening 17,008 articles, 36 studies, including 10 pilot studies, were included (n = 5,165 ICU patients). Interventions were subdivided into six categories: 1) exercise and physical rehabilitation programs; 2) follow-up services; 3) psychosocial programs; 4) diaries; 5) information and education; and 6) other interventions. Many outcomes favored the interventions, but significant differences were only found for diaries in reducing depression (two studies, n = 88; standardized mean difference, 0.68; 95% CI, 0.14–1.21) and anxiety (two studies, n = 88; standardized mean difference, 0.44; 95% CI, 0.01–0.87) and exercise programs in improving the Short Form Health Survey-36 Mental Component Score (seven studies, n = 664; mean difference, 2.62; 95% CI, 0.92–4.32). Conclusions: There is thin evidence that diaries and exercise programs have a positive effective on mental outcomes. Despite outcomes favoring the intervention group, other commonly used nonpharmacologic interventions in daily ICU practice are not supported by conclusive evidence from this meta-analysis. To improve recovery programs for ICU survivors, more evidence is needed from robust intervention studies using standardized outcomes. Registration: International prospective register of systematic reviews (PROSPERO) NCT01738620. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: wytske.geense@radboudumc.nl Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Global and Regional Diagnostic Accuracy of Lung Ultrasound Compared to CT in Patients With Acute Respiratory Distress Syndrome Objectives: Lung CT is the reference imaging technique for acute respiratory distress syndrome, but requires transportation outside the intensive care and x-ray exposure. Lung ultrasound is a promising, inexpensive, radiation-free, tool for bedside imaging. Aim of the present study was to compare the global and regional diagnostic accuracy of lung ultrasound and CT scan. Design: A prospective, observational study. Setting: Intensive care and radiology departments of a University hospital. Patients: Thirty-two sedated, paralyzed acute respiratory distress syndrome patients (age 65 ± 14 yr, body mass index 25.9 ± 6.5 kg/m2, and PaO2/FIO2 139 ± 47). Interventions: Lung CT scan and lung ultrasound were performed at positive end-expiratory pressure 5 cm H2O. A standardized assessment of six regions per hemithorax was used; each region was classified for the presence of normal aeration, alveolar-interstitial syndrome, consolidation, and pleural effusion. Agreement between the two techniques was calculated, and diagnostic variables were assessed for lung ultrasound using lung CT as a reference. Measurements and Main Results: Global agreement between lung ultrasound and CT ranged from 0.640 (0.391–0.889) to 0.934 (0.605–1.000) and was on average 0.775 (0.577–0.973). The overall sensitivity and specificity of lung ultrasound ranged from 82.7% to 92.3% and from 90.2% to 98.6%, respectively. Similar results were found with regional analysis. The diagnostic accuracy of lung ultrasound was significantly higher when those patterns not reaching the pleural surface were excluded (area under the receiver operating characteristic curve: alveolar-interstitial syndrome 0.854 [0.821–0.887] vs 0.903 [0.852–0.954]; p = 0.049 and consolidation 0.851 [0.818–0.884] vs 0.896 [0.862–0.929]; p = 0.044). Conclusions: Lung ultrasound is a reproducible, sensitive, and specific tool, which allows for bedside detections of the morphologic patterns in acute respiratory distress syndrome. The presence of deep lung alterations may impact the diagnostic performance of this technique. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: chiumello@libero.it Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Quantifiable Bleeding in Children Supported by Extracorporeal Membrane Oxygenation and Outcome Objectives: To evaluate the association between bleeding from chest tubes and clinical outcomes in children supported by extracorporeal membrane oxygenation. Design: Secondary analysis of a large observational cohort study. Setting: Eight pediatric institutions within the Eunice Kennedy Shriver National Institute of Child Health and Human Development’s Collaborative Pediatric Critical Care Research Network. Patients: Critically ill children supported by extracorporeal membrane oxygenation. Interventions: None. Measurements and Main Results: Daily chest tube bleeding, results from hemostatic assays, transfusion volumes, 90-day PICU-free days, and mortality were collected prospectively by trained bedside extracorporeal membrane oxygenation specialists and research coordinators. Extracorporeal membrane oxygenation was employed in the care of 514 consecutive patients. Sixty percent of patients had at least one episode of chest tube bleeding (median chest tube blood volume over the entire extracorporeal membrane oxygenation course was 123 mL/kg [interquartile range, 47–319 mL/kg]). Twenty-six percent had at least 1 day of bleeding from the chest tube greater than 100 mL/kg/d. The number of days with chest tube bleeding greater than 60 mL/kg/d was independently associated with increased in-hospital mortality (adjusted odds ratio, 1.43; 95% CI, 1.05–1.97; p = 0.02) and decreased PICU-free days (beta coefficient, –4.2; 95% CI, –7.7 to –0.6; p = 0.02). The total amount of bleeding from chest tube were independently associated with increased mortality (per mL/kg/extracorporeal membrane oxygenation run; adjusted odds ratio, 1.002; 95% CI, 1.000–1.003; p = 0.04). Fibrinogen, weight, indication for extracorporeal membrane oxygenation, and need for hemodialysis were independently associated with chest tube bleeding, whereas platelet count, coagulation tests, heparin dose, and thrombotic events were not. Conclusions: In children supported by extracorporeal membrane oxygenation, chest tube bleeding above 60 mL/kg/d was independently associated with worse clinical outcome. Low fibrinogen was independently associated with chest tube bleeding, whereas platelet count and hemostatic tests were not. Further research is needed to evaluate if interventions to prevent or stop chest tube bleeding influence the clinical outcome. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). The Collaborative Pediatric Critical Care Research Network was funded by the National Institute of Child Health and Human Development (NICHD). This article does not necessarily reflect the opinions or views of the Bleeding and Thrombosis During Extracorporeal Membrane Oxygenation Investigators or the NICHD. Dr. Dalton received funding from Innovative Extracorporeal Membrane Oxygenation Concepts, and she received support for article research from the National Institutes of Health. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: man9026@med.cornell.edu Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.