Social Media Usage for Medical Education and Smartphone Addiction Among Medical Students: National Web-Based Survey.

Background: Social media (SoMe) have taken a major place in the medical field, and younger generations are increasingly using them as their primary source to find information. Objective: This study aimed to describe the use of SoMe for medical education among French medical students and assess the prevalence of smartphone addiction in this population. Methods: A cross-sectional web-based survey was conducted among French medical students (second to sixth year of study). The questionnaire collected information on SoMe use for medical education and professional behavior. Smartphone addiction was assessed using the Smartphone Addiction Scale Short-Version (SAS-SV) score. Results: A total of 762 medical students responded to the survey. Of these, 762 (100%) were SoMe users, spending a median of 120 (IQR 60-150) minutes per day on SoMe; 656 (86.1%) used SoMe for medical education, with YouTube, Instagram, and Facebook being the most popular platforms. The misuse of SoMe in a professional context was also identified; 27.2% (207/762) of students posted hospital internship content, and 10.8% (82/762) searched for a patient's name on SoMe. Smartphone addiction was prevalent among 29.1% (222/762) of respondents, with a significant correlation between increased SoMe use and SAS-SV score (r=0.39, 95% CI 0.33-0.45; P<.001). Smartphone-addicted students reported a higher impact on study time (211/222, 95% vs 344/540, 63.6%; P<.001) and a greater tendency to share hospital internship content on social networks (78/222, 35.1% vs 129/540, 23.8%; P=.002). Conclusions: Our findings reveal the extensive use of SoMe for medical education among French medical students, alongside a notable prevalence of smartphone addiction. These results highlight the need for medical schools and educators to address the responsible use of SoMe and develop strategies to mitigate the risks associated with excessive use and addiction.

Barriers encountered with clinical data warehouses: Recommendations from a focus group.

BACKGROUND AND OBJECTIVE: The increasing implementation and use of electronic health records over the last few decades has made a significant volume of clinical data being available. Over the past 20 years, hospitals have also adopted and implemented data warehouse technology to facilitate the reuse of administrative and clinical data for research. However, the implementation of clinical data warehouses encounters a set of barriers: ethical, legislative, technical, human and organizational. This paper proposes an overview of difficulties and barriers encountered during a clinical data warehouse (CDW) development and implementation project. METHODS: We conducted a focus group at the 2023 Medical Informatics Europe Conference and invited professionals involved in the implementation of CDW. These experts described their CDW and the difficulties and barriers they encountered at each phase: (i) launching of the data warehouse project, (ii) implementing the data warehouse and (iii) using a data warehouse in routine operations. They were also asked to propose solutions they were able to implement to address the barriers previously reported. RESULTS: After synthesis and consensus, a total of 26 barriers were identified, 10 pertained to tasks, 5 to tools and technologies, 4 to persons, 4 to organization, and 3 to the external environment. To address these challenges, a set of 15 practical recommendations was offered, covering essential aspects such as governance, stakeholder engagement, interdisciplinary collaboration, and external expertise utilization. CONCLUSIONS: These recommendations serve as a valuable resource for healthcare institutions seeking to establish and optimize CDWs, offering a roadmap for leveraging clinical data for research, quality enhancement, and improved patient care.

How to Accurately Detect Renal Replacement Therapy Weaning in Intensive Care: Data Quality and Standardization Considerations for the OMOP Common Data Model.

This study assesses the effectiveness of the Observational Medical Outcomes Partnership common data model (OMOP CDM) in standardising Continuous Renal Replacement Therapy (CRRT) data from intensive care units (ICU) of two French university hospitals. Our objective was to extract and standardise data from various sources, enabling the development of predictive models for CRRT weaning that are agnostic to the data's origin. Data for 1,696 ICU stays from the two data sources were extracted, transformed, and loaded into the OMOP format after semantic alignment of 46 CRRT standard concepts. Although the OMOP CDM demonstrated potential in harmonising CRRT data, we encountered challenges related to data variability and the lack of standard concepts. Despite these challenges, our study supports the promise of the OMOP CDM for ICU data standardization, suggesting that further refinement and adaptation could significantly improve clinical decision making and patient outcomes in critical care settings.

Intensive Care Quality Indicators Dashboard Using Observational Medical Outcomes Partnership Common Data Model.

This paper presents the development of a visualization dashboard for quality indicators in intensive care units (ICUs), using the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). The dashboard enables the user to visualize quality indicator data using histograms, pie charts and tables. Our project uses the OMOP CDM, ensuring a seamless implementation of our dashboard across various hospitals. Future directions for our research include expanding the dashboard to incorporate additional quality indicators and evaluating clinicians' feedback on its effectiveness.

Public Availability of Data Management and Analysis Scripts of Studies Conducted on Open-Access Intensive Care Databases.

Intensive care units (ICUs) provide care for critical patients at high risk of morbidity and mortality, and require continuous monitoring of clinical, biological and, imaging parameters. Collaborative ventures have enabled the emergence of large open access databases for the secondary use of Electronic Health Records (EHRs). The objective of this work was to evaluate the availability of scripts and datasets in publications based on ICU open-access databases. We included 910 original articles based on four ICU open-access databases (Amsterdam University Medical Centers Database, eICU Collaborative Research Database, High time resolution ICU dataset, and Medical Information Mart for Intensive Care). The majority of the studies did not provide their data management scripts (n=839, 92.9%), neither the analysis script (n=843, 93.4%) in the article. Attempts to contact the 845 corresponding authors in question resulted in 89.11% (n=753) of our e-mail requests going unanswered over a two-month period. We received 51 automated messages (55.43%) indicating that emails have not been delivered, while 6 messages (6.52%) redirected to alternative email addresses. Only 20 corresponding authors (18.18%) answered, finally providing the requested materials. Despite scientific journals recommendations to share materials, our study unveils the absence of crucial components for the replication of studies by other research teams.

Prognosis of major trauma in patients older than 85 years admitted to the ICU, a registry-based study.

BACKGROUND: The aging population in France and Western Europe is on the rise, particularly among individuals aged 65 years and older. Although older adults are susceptible to traumatic injuries, they constitute a minority of trauma center admissions especially those aged 85 and above. The aim of our study was to investigate the prognostic factors for mortality among the older old population (aged 85 years and above) managed in ICU of Traumabase group trauma centers. METHODS: This retrospective observational cohort study, conducted from 2013 to 2022, analyzed all severely injured older patients (aged ≥ 85 years) managed in 14 ICU trauma centers enrolled in the Traumabase registry. The study examined sociodemographic, clinical, and outcome variables. Frailty was assessed using the Clinical Frailty Scale. RESULTS: Among the 365 older trauma patients, 190 (52.1%) were classified as non-frail (CFS 1-3), 80 (21.9%) as pre-frail (CFS 4,5), and 95 (26%) as frail (CFS 6-9). Falls were the most common mechanism of injury. High mortality rates were observed, with 43.5% ICU mortality and 45.5% mortality at day 30. Factors most associated with ICU mortality included traumatic brain injury (CGS < 13), pre-hospital micromethod hemoglobin < 13 and severity of injury (ISS > 16). CONCLUSION: Factors such as traumatic brain injury and severe hemorrhage (micromethod hemoglobin < 13) and ISS > 16 are associated with ICU mortality in in patients older than 85 years trauma patient. Early geriatric intervention is crucial for optimizing outcomes in this vulnerable population.

Comparison of Two Dexmedetomidine Administration Strategies on the Incidence of Postoperative Respiratory Complications: A Retrospective, Inverse Probability of Treatment Weighted Study.

Randomized controlled trials have shown a higher risk of postoperative hypoxemia and delayed extubation with opioid-free anesthesia (OFA), compared with opioid anesthesia. The practice of OFA is not standardized. The objective of this study is to investigate the association between the dexmedetomidine administration protocol used and the occurrence of postoperative respiratory complications. This work is a retrospective, propensity score-adjusted study (inverse probability of treatment weighting) conducted between January 2019 and September 2021 in a French tertiary care university hospital, including 180 adult patients undergoing major digestive surgery. Comparison of 2 anesthesia protocols: with a continuous intravenous maintenance dose of dexmedetomidine following a bolus (group B+M, n = 105) or with a bolus dose alone (group B, n = 75). The main outcome measure was a composite respiratory end point within 24 hours of surgery. There was no significant difference in the incidence of overall respiratory complications, as assessed by the primary end point. Nevertheless, there were more patients with postoperative hypercapnia in group B+M than in group B (16% vs 2.5%, P = .004). Patients in group B+M were extubated later than patients in group B (group B+M, median 40 minutes, IQR 20-74 minutes; group B, median 20 minutes, IQR 10-50 minutes; P = .004). Our study showed negative results for the primary end point. However, data on the increased risk of postoperative hypercapnia in patients receiving a maintenance dose of dexmedetomidine are new. Other prospective randomized studies with greater power are necessary to confirm these data and to make OFA safer, by reducing the prescribed doses of dexmedetomidine.

Effect of opioid-free anesthesia on post-operative period in cytoreductive surgery with hyperthermic intraperitoneal chemotherapy: a propensity score matched study.

BACKGROUND: Postoperative pain after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS-HIPEC) is important. It appears essential to reduce postoperative pain and morphine consumption. METHODS: Retrospective study in a university hospital comparing patient benefiting from CRS-HIPEC under opioid-free anesthesia (OFA; dexmedetomidine) to those anesthetized with opioid anesthesia (OA; remifentanil) using a propensity score matching method. The main objective was the impact of OFA on postoperative morphine consumption in the first 24 h after surgery. RESULTS: 102 patients were included, matching on the propensity score allowed selecting 34 unique pairs analyzed. Morphine consumption was lower in the OFA group than in the OA group (3.0 [0.00-11.0] mg/24 h vs. 13.0 [2.5-25.0] mg/24 h; p = 0.02). In multivariable analysis, OFA was associated with a reduction of 7.2 [0.5-13.9] mg of postoperative morphine (p = 0.04). The rate of renal failure with a KDIGO-score > 1 was lower in the OFA group than in the OA group (12% vs. 38%; p = 0.01). There was no difference between groups concerning length of surgery/anesthesia, norepinephrine infusion, volume of fluid therapy, post-operative complications, rehospitalization or ICU readmission within 90 days, mortality, and postoperative rehabilitation. CONCLUSION: Our results suggest that OFA for CRS-HIPEC patients appears safe and is associated with less postoperative morphine use and acute kidney injury.

Evaluation of Medical Confidentiality Breaches on Twitter Among Anesthesiology and Intensive Care Health Care Workers.

BACKGROUND: With the generalization of social network use by health care workers, we observe the emergence of breaches in medical confidentiality. Our objective was to determine, among anesthesiology and intensive care health care workers, the rate of medical confidentiality breaches among professional tweets. METHODS: We performed a retrospective analysis of public Twitter data available through the official Twitter application program interface. The profiles of anesthesiology and intensive care professionals were identified thanks to keywords in their biography. All the tweets with a photograph and all the text-only tweets containing at least one specific keyword related to anesthesiology or intensive care were extracted. We selected only the tweets with a health care-related character. Then, we analyzed 10% of the tweets with a photograph and 10% of the text-only tweets extracted and noted those presenting a breach of medical confidentiality. RESULTS: After a first screening of 12,705 accounts, we manually analyzed 431 tweets with photograph(s) and 9000 text-only tweets from 1831 accounts. We found 44 (10.2%) breaches of medical confidentiality among the photographs and 76 (0.8%) among text-only tweets. These 120 problematic tweets came from 96 profiles (96/1831; 5.2%); 3.7% of North American profiles breached medical confidentiality versus 6.3% of profiles from other areas; P = .03. When comparing the distribution of the number of followers and tweets, accounts with breach of medical confidentiality tweets had more tweets and followers than profiles without (both P < .0001). CONCLUSIONS: We found a significant proportion of tweets with breach of medical confidentiality among anesthesiology and intensive care professionals accounts.

Trends in major intensive care medicine journals: A machine learning approach.

PURPOSE: Intensive care medicine (ICM) has the particularity of being a multidisciplinary specialty and its literature reflects this multidisciplinarity. However, the proportion of each field in this literature and its trend dynamics are not known. The objective of this study was to analyze the ICM literature, extract latent topics and search for the presence of research trends. MATERIAL AND METHODS: Abstracts of original articles from the top ICM journals, from their inception until December 31st, 2019, were included. This corpus was fed into a structural topic modeling algorithm to extract latent semantic topics. The temporal distribution was then analyzed and the presence of trends was searched by Mann-Kendall trends tests. RESULTS: Finally, 49,276 articles from 10 journals were included. After topic modeling analysis and experts' feedback, 124 research topics were selected and labeled. Topics were categorized into 19 categories, the most represented being respiratory, fundamental and neurological research. Increasing trends were observed for research on mechanical ventilation and decreasing trends for cardiopulmonary resuscitation. CONCLUSIONS: This study reviewed all articles from major ICM journals in a comprehensive way. It provides a better understanding of ICM research landscape by analyzing the temporal evolution of latent research topics in the ICM literature.

Standardized Description of the Feature Extraction Process to Transform Raw Data Into Meaningful Information for Enhancing Data Reuse: Consensus Study.

BACKGROUND: Despite the many opportunities data reuse offers, its implementation presents many difficulties, and raw data cannot be reused directly. Information is not always directly available in the source database and needs to be computed afterwards with raw data for defining an algorithm. OBJECTIVE: The main purpose of this article is to present a standardized description of the steps and transformations required during the feature extraction process when conducting retrospective observational studies. A secondary objective is to identify how the features could be stored in the schema of a data warehouse. METHODS: This study involved the following 3 main steps: (1) the collection of relevant study cases related to feature extraction and based on the automatic and secondary use of data; (2) the standardized description of raw data, steps, and transformations, which were common to the study cases; and (3) the identification of an appropriate table to store the features in the Observation Medical Outcomes Partnership (OMOP) common data model (CDM). RESULTS: We interviewed 10 researchers from 3 French university hospitals and a national institution, who were involved in 8 retrospective and observational studies. Based on these studies, 2 states (track and feature) and 2 transformations (track definition and track aggregation) emerged. "Track" is a time-dependent signal or period of interest, defined by a statistical unit, a value, and 2 milestones (a start event and an end event). "Feature" is time-independent high-level information with dimensionality identical to the statistical unit of the study, defined by a label and a value. The time dimension has become implicit in the value or name of the variable. We propose the 2 tables "TRACK" and "FEATURE" to store variables obtained in feature extraction and extend the OMOP CDM. CONCLUSIONS: We propose a standardized description of the feature extraction process. The process combined the 2 steps of track definition and track aggregation. By dividing the feature extraction into these 2 steps, difficulty was managed during track definition. The standardization of tracks requires great expertise with regard to the data, but allows the application of an infinite number of complex transformations. On the contrary, track aggregation is a very simple operation with a finite number of possibilities. A complete description of these steps could enhance the reproducibility of retrospective studies.

Transforming Anesthesia Data Into the Observational Medical Outcomes Partnership Common Data Model: Development and Usability Study.

BACKGROUND: Electronic health records (EHRs, such as those created by an anesthesia management system) generate a large amount of data that can notably be reused for clinical audits and scientific research. The sharing of these data and tools is generally affected by the lack of system interoperability. To overcome these issues, Observational Health Data Sciences and Informatics (OHDSI) developed the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) to standardize EHR data and promote large-scale observational and longitudinal research. Anesthesia data have not previously been mapped into the OMOP CDM. OBJECTIVE: The primary objective was to transform anesthesia data into the OMOP CDM. The secondary objective was to provide vocabularies, queries, and dashboards that might promote the exploitation and sharing of anesthesia data through the CDM. METHODS: Using our local anesthesia data warehouse, a group of 5 experts from 5 different medical centers identified local concepts related to anesthesia. The concepts were then matched with standard concepts in the OHDSI vocabularies. We performed structural mapping between the design of our local anesthesia data warehouse and the OMOP CDM tables and fields. To validate the implementation of anesthesia data into the OMOP CDM, we developed a set of queries and dashboards. RESULTS: We identified 522 concepts related to anesthesia care. They were classified as demographics, units, measurements, operating room steps, drugs, periods of interest, and features. After semantic mapping, 353 (67.7%) of these anesthesia concepts were mapped to OHDSI concepts. Further, 169 (32.3%) concepts related to periods and features were added to the OHDSI vocabularies. Then, 8 OMOP CDM tables were implemented with anesthesia data and 2 new tables (EPISODE and FEATURE) were added to store secondarily computed data. We integrated data from 5,72,609 operations and provided the code for a set of 8 queries and 4 dashboards related to anesthesia care. CONCLUSIONS: Generic data concerning demographics, drugs, units, measurements, and operating room steps were already available in OHDSI vocabularies. However, most of the intraoperative concepts (the duration of specific steps, an episode of hypotension, etc) were not present in OHDSI vocabularies. The OMOP mapping provided here enables anesthesia data reuse.

Effect of early hyperoxemia on mortality in mechanically ventilated septic shock patients according to Sepsis-3 criteria: analysis of the MIMIC-III database.

BACKGROUND AND IMPORTANCE: Hyperoxemia may be associated with increased mortality in emergency room or ICU patients. However, its effect during septic shock is still debated. OBJECTIVE: To evaluate the effect of hyperoxemia on ICU mortality, during the first 24 h of ICU stay, in mechanically ventilated patients with septic shock according to SEPSIS-3 criteria. DESIGN, SETTINGS AND PARTICIPANTS: A retrospective cohort study of ICU admissions recorded in the Medical Information Mart for Intensive Care-III, a retrospective ICU database, was performed. INTERVENTION: Two oxygen exposures during the first 24 h were compared: average PaO2 (TWA-PaO2) between 70 and 120 mmHg in the normoxemia group and above 120 mmHg in the hyperoxemia group. OUTCOME MEASURES AND ANALYSIS: The primary outcome was mortality during ICU stay. MAIN RESULTS: Four hundred eighty-eight ICU admissions met the inclusion criteria: 214 in the normoxemia group and 274 in the hyperoxemia group. The median TWA-PaO2 was 99.1 (88.9-107.6) mmHg in the normoxemia group and 151.5 (133.6-180.2) mmHg in the hyperoxemia group. ICU mortality was lower in the hyperoxemia group than in the normoxemia group in univariate analysis [29.6 vs. 39.7%, respectively; OR 0.64 (0.44-0.93); P = 0.024], but not in multivariate analysis [OR 0.98 (0.62-1.56); P = 0.93]. There was no difference between the two groups in ICU length of stay [8.0 (4.3-15.0) vs. 8.4 (4.7-15.0) days; P = 0.82]. CONCLUSION: We did not find any impact of early hyperoxemia on mortality in this population of mechanically ventilated patients with SEPSIS-3 septic shock criteria.

Association of Social Network Use With Increased Anxiety Related to the COVID-19 Pandemic in Anesthesiology, Intensive Care, and Emergency Medicine Teams: Cross-Sectional Web-Based Survey Study.

BACKGROUND: Critical care teams are on the front line of managing the COVID-19 pandemic, which is stressful for members of these teams. OBJECTIVE: Our objective was to assess whether the use of social networks is associated with increased anxiety related to the COVID-19 pandemic among members of critical care teams. METHODS: We distributed a web-based survey to physicians, residents, registered and auxiliary nurses, and nurse anesthetists providing critical care (anesthesiology, intensive care, or emergency medicine) in several French hospitals. The survey evaluated the respondents' use of social networks, their sources of information on COVID-19, and their levels of anxiety and information regarding COVID-19 on analog scales from 0 to 10. RESULTS: We included 641 respondents in the final analysis; 553 (86.3%) used social networks, spending a median time of 60 minutes (IQR 30-90) per day on these networks. COVID-19-related anxiety was higher in social network users than in health care workers who did not use these networks (median 6, IQR 5-8 vs median 5, IQR 3-7) in univariate (P=.02) and multivariate (P<.001) analyses, with an average anxiety increase of 10% in social network users. Anxiety was higher among health care workers using social networks to obtain information on COVID-19 than among those using other sources (median 6, IQR 5-8 vs median 6, IQR 4-7; P=.04). Social network users considered that they were less informed about COVID-19 than those who did not use social networks (median 8, IQR 7-9 vs median 7, IQR 6-8; P<.01). CONCLUSIONS: Our results suggest that social networks contribute to increased anxiety in critical care teams. To protect their mental health, critical care professionals should consider limiting their use of these networks during the COVID-19 pandemic.

Correction to: Early sepsis markers in patients admitted to intensive care unit with moderate­to­severe diabetic ketoacidosis.

An amendment to this paper has been published and can be accessed via the original article.

Early sepsis markers in patients admitted to intensive care unit with moderate-to-severe diabetic ketoacidosis.

BACKGROUND: Bacterial infections are frequent triggers for diabetic ketoacidosis. In this context, delayed antibiotic treatment is associated with increased morbidity and mortality. Unnecessary administration of antimicrobial therapy might however, also negatively impact the prognosis. The usefulness of sepsis markers in diabetic ketoacidosis has not been assessed. Thus, we sought to investigate diagnostic performances of clinical and biological sepsis markers during diabetic ketoacidosis. METHODS: In this monocentric retrospective cohort study, all consecutive episodes of diabetic ketoacidosis (defined as pH ≤ 7.25, glycaemia > 300 mg/dL and presence of ketones) admitted in intensive care unit were included. A proven bacterial infection was defined as bacteriological documentation on any bacterial sample. Clinical (presence of fever: temperature > 38 °C and presence of hypothermia: temperature < 36 °C) and biological markers (whole blood count, neutrophils count, neutrophils-to-lymphocytes count ratio and procalcitonin), recorded at admission, were compared according to the presence or absence of a proven bacterial infection. RESULTS: Between 2011 and 2018, among 134 episodes of diabetic ketoacidosis, 102 were included (91 patients). Twenty out of 102 were infected. At admission, procalcitonin (median: 3.58 ng/mL vs 0.52 ng/mL, p < 0.001) and presence of fever (25% vs 44%, p = 0.007) were different between episodes with and without proven bacterial infection in both univariate and multivariate analysis. Whole blood count, neutrophils count, neutrophils-to-lymphocytes count ratio and presence of hypothermia were not different between both groups. The diagnostic performance analysis for procalcitonin revealed an area under the curve of 0.87 with an optimal cutoff of 1.44 ng/mL leading to a sensitivity of 0.90 and a specificity of 0.76. Combining procalcitonin and presence of fever allowed to distinguish proven bacterial infection episodes from those without proven bacterial infection. Indeed, all patients with procalcitonin level of more than 1.44 ng/mL and fever had proven bacterial infection episodes. The presence of one of these 2 markers was associated with 46% of proven bacterial infection episodes. No afebrile patient with procalcitonin level less than 1.44 ng/mL had a proven bacterial infection. CONCLUSION: At admission, combining procalcitonin and presence of fever may be of value to distinguish ketoacidosis patients with and without proven bacterial infection, admitted in intensive care unit.

Early Positive Fluid Balance is Associated with Mortality in Patients Treated with Veno-Arterial Extra Corporeal Membrane Oxygenation For Cardiogenic Shock: a Retrospective Cohort Study.

PURPOSE: Veno-arterial Extracoporeal Membrane Oxygenation (VA-ECMO) provides circulatory assistance for patients in cardiogenic shock. Large quantities of fluids are often required, especially during the early stages, but can result in a potentially harmful fluid-overload. The objective was to determine the association of early fluid-balance and mortality. METHODS: Retrospective single-center study in 101 VA-ECMO patients between 2013 and 2016. Daily fluid-balance and weight changes over the 5 first days were observed. Analyses between survivors and non-survivors were conducted using Mann-Whitney tests and logistic regression multivariable and Cox hazard-proportional analyses to determine any association with 28-days mortality. RESULTS: Mortality was 47.5%. The fluid-balance was higher in non-survivors at day-1 (47.3[18.1-71.9] vs. 19.3[1.5-36.2] mL/kg, P < 0.0001) and day-2 (30.6[14.8-71.0] vs. 10.1[-9.8 to 34.7] mL/kg, P = 0.025), as was the cumulative fluid-balance over the first 5 days (107.3[40.5-146.2] vs. 53.0[7.5-74.3] mL/kg, P = 0.04). The administration of unintentional fluids (used for preparation and infusion of drugs) represented an important part of the administrated fluids (15 mL/kg/d-23 mL/kg/d). A significant but moderate correlation was observed between fluid-balance and weight variations over the 5 days (r values ranging from 0.36 to 0.54). Among other parameters, day-1 fluid-balance was independently associated with mortality (OR = 14.34 [1.58-129.79], P = 0.02) and day-1 and day-2 with time to death (HR = 8.26 [1.12-60.98], P = 0.04 and 2.89 [1.26-6.65], P = 0.01). A threshold of 38.8 mL/kg predicted mortality with a sensitivity of 60% and specificity of 83% (area under the curve: 0.749). CONCLUSION: Early positive fluid-balance is associated with mortality in VA-ECMO patients.