Evaluation of ChatGPT in Predicting 6-Month Outcomes After Traumatic Brain Injury.

OBJECTIVES: To evaluate the capacity of ChatGPT, a widely accessible and uniquely popular artificial intelligence-based chatbot, in predicting the 6-month outcome following moderate-to-severe traumatic brain injury (TBI). DESIGN: Single-center observational retrospective study. SETTING: Data are from a neuro-ICU from a level 1 trauma center. PATIENTS: All TBI patients admitted to ICU between September 2021 and October 2022 were included in a prospective database. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Based on anonymized clinical, imaging, and biological information available at the patients' hospital admission and extracted from the database, clinical vignettes were retrospectively submitted to ChatGPT for prediction of patients' outcomes. The predictions of two intensivists (one neurointensivist and one non-neurointensivist) both from another level 1 trauma center (Beaujon Hospital), were also collected as was the International Mission on Prognosis and Analysis of Clinical Trials in Traumatic Brain Injury (IMPACT) scoring. Each intensivist, as well as ChatGPT, made their prognostic evaluations independently, without knowledge of the others' predictions and of the patients' actual management and outcome. Both the intensivists and ChatGPT were given access to the exact same set of information. The main outcome was a 6-month-functional status dichotomized into favorable (Glasgow Outcome Scale Extended [GOSE] ≥ 5) versus poor (GOSE < 5). Prediction of intracranial hypertension management, pulmonary infectious risk, and removal of life-sustaining therapies was also investigated as secondary outcomes. Eighty consecutive moderate-to-severe TBI patients were included. For the 6-month outcome prognosis, area under the receiver operating characteristic curve (AUC-ROC) for ChatGPT, the neurointensivist, the non-neurointensivist, and IMPACT were, respectively, 0.62 (0.50-0.74), 0.70 (0.59-0.82), 0.71 (0.59-0.82), and 0.81 (0.72-0.91). ChatGPT had the highest sensitivity (100%), but the lowest specificity (26%). For secondary outcomes, ChatGPT's prognoses were generally less accurate than clinicians' prognoses, with lower AUC values for most outcomes. CONCLUSIONS: This study does not support the use of ChatGPT for prediction of outcomes after TBI.

Evaluation of transcranial Doppler use in patients with acute liver failure listed for emergency liver transplantation.

PURPOSE: Acute liver failure (ALF) is characterized by hepatic encephalopathy (HE) often due to intracranial hypertension (ICH). The risk/benefit-balance of intraparenchymal pressure catheter monitoring is controversial during ALF. AIMS: Perform an evaluation of transcranial Doppler (TCD) use in patients with ALF listed for emergency liver transplantation. MATERIAL AND METHODS: Single center retrospective cohort study including all patients registered on high emergency LT list between 2012 and 2018. All TCD measurements performed during ICU stay after listing and after LT (when performed) were recorded. TCD was considered abnormal when pulsatility index (PI) was >1.2. RESULTS: Among 106 patients with ALF, forty-seven (44%) had a TCD while on list. They had more severe liver and extrahepatic organ failure. When performed, TCD was abnormal in 51% of patients. These patients more frequently developed ICH events (45% vs. 13%, p = .02) and more frequently required increase in sedative drugs and vasopressors. While 22% of patients with normal TCD spontaneously survived, all of those with abnormal TCD died or were transplanted (p = .02). All transplanted patients who had abnormal exams normalized their TCD within 2 (1-2) days after LT. CONCLUSION: TCD may be a useful non-invasive tool for ICH detection and management, then guide sedation withdrawal.

Limitation of life-sustaining therapies in critically ill patients with COVID-19: a descriptive epidemiological investigation from the COVID-ICU study.

BACKGROUND: Limitations of life-sustaining therapies (LST) practices are frequent and vary among intensive care units (ICUs). However, scarce data were available during the COVID-19 pandemic when ICUs were under intense pressure. We aimed to investigate the prevalence, cumulative incidence, timing, modalities, and factors associated with LST decisions in critically ill COVID-19 patients. METHODS: We did an ancillary analysis of the European multicentre COVID-ICU study, which collected data from 163 ICUs in France, Belgium and Switzerland. ICU load, a parameter reflecting stress on ICU capacities, was calculated at the patient level using daily ICU bed occupancy data from official country epidemiological reports. Mixed effects logistic regression was used to assess the association of variables with LST limitation decisions. RESULTS: Among 4671 severe COVID-19 patients admitted from February 25 to May 4, 2020, the prevalence of in-ICU LST limitations was 14.5%, with a nearly six-fold variability between centres. Overall 28-day cumulative incidence of LST limitations was 12.4%, which occurred at a median of 8 days (3-21). Median ICU load at the patient level was 126%. Age, clinical frailty scale score, and respiratory severity were associated with LST limitations, while ICU load was not. In-ICU death occurred in 74% and 95% of patients, respectively, after LST withholding and withdrawal, while median survival time was 3 days (1-11) after LST limitations. CONCLUSIONS: In this study, LST limitations frequently preceded death, with a major impact on time of death. In contrast to ICU load, older age, frailty, and the severity of respiratory failure during the first 24 h were the main factors associated with decisions of LST limitations.

Continuous and deep sedation until death after a decision to withdraw life-sustaining therapies in intensive care units: A national survey.

BACKGROUND: Continuous and deep sedation until death is a much highly debated end-of-life practice. France is unique in having a regulatory framework for it. However, there are no data on its practice in intensive care units (ICUs). AIM: The aim is to describe continuous and deep sedation in relation to the framework in the specific context of withdrawal of life-sustaining therapies in ICUs, that is, its decision-making process and its practice compared to other end-of-life practices in this setting. DESIGN AND SETTING: French multicenter observational study. Consecutive ICU patients who died after a decision to withdraw life-sustaining therapies. RESULTS: A total of 343 patients in 57 ICUs, 208 (60%) with continuous and deep sedation. A formalized procedure for continuous and deep sedation was available in 32% of the ICUs. Continuous and deep sedation was not the result of a collegial decision-making process in 17% of cases, and did not involve consultation with an external physician in 29% of cases. The most commonly used sedative medicines were midazolam (10 [5-18] mg h-1) and propofol (200 [120-250] mg h -1). The Richmond Agitation Sedation Scale (RASS) was -5 in 60% of cases. Analgesia was associated with sedation in 94% of cases. Compared with other end-of-life sedative practices (n = 98), medicines doses were higher with no difference in the depth of sedation. CONCLUSIONS: This study shows a poor compliance with the framework for continuous and deep sedation. It highlights the need to formalize it to improve the decision-making process and the match between the intent, the practice and the actual effect.

Type I interferon signaling in systemic immune cells from patients with alcoholic cirrhosis and its association with outcome.

BACKGROUND & AIMS: In immune cells, constitutively and acutely produced type I interferons (IFNs) engage autocrine/paracrine signaling pathways to induce IFN-stimulated genes (ISGs). Enhanced activity of IFN signaling pathways can cause excessive inflammation and tissue damage. We aimed to investigate ISG expression in systemic immune cells from patients with decompensated alcoholic cirrhosis, and its association with outcome. METHODS: Peripheral blood mononuclear cells (PBMCs) from patients and heathy subjects were stimulated or not with lipopolysaccharide (LPS, an IFN inducer) or increasing concentrations of IFN-ß. The expression of 48 ISGs and ten "non-ISG" inflammatory cytokines were analyzed using RT-qPCR. RESULTS: We developed an 8-ISG signature (IFN score) assessing ISG expression. LPS-stimulated ISG induction was significantly lower in PBMCs from patients with cirrhosis compared to healthy controls. Non-ISGs, however, showed higher induction. Lower induction of ISGs by LPS was not due to decreased IFN production by cirrhotic PBMCs or neutralization of secreted IFN, but a defective PBMC response to IFN. This defect was at least in part due to decreased constitutive ISG expression. Patients with the higher baseline IFN scores and ISG levels had the higher risk of death. At baseline, "non-ISG" cytokines did not correlate with outcome. CONCLUSIONS: PBMCs from patients with decompensated alcoholic cirrhosis exhibit downregulated ISG expression, both constitutively and after an acute stimulus. Our finding that higher baseline PBMC ISG expression was associated with higher risk of death, suggests that constitutive ISG expression in systemic immune cells contributes to the prognosis of alcoholic cirrhosis. LAY SUMMARY: Enhanced activity of IFN signaling pathways can cause excessive inflammation and tissue damage. Here we show that peripheral blood mononuclear cells (PBMCs) from patients with alcoholic cirrhosis exhibit a defect in interferon-stimulated genes (ISGs). We found that higher baseline ISG expression in PBMCs was associated with higher risk of death, revealing a probable contribution of ISG expression in immune cells to the outcome of alcoholic cirrhosis.

Tension between continuous and deep sedation and assistance in dying: a national survey of intensive care professionals' perceptions.

INTRODUCTION: The situation in France is unique, having a legal framework for continuous and deep sedation (CDS). However, its use in intensive care units (ICU), combined with the withdrawal of life-sustaining therapies, still raises ethical issues, particularly its potential to hasten death. The legalization of assistance in dying, i.e., assisted suicide or euthanasia at the patient's request, is currently under discussion in France. The objectives of this national survey were first, to assess whether ICU professionals perceive CDS administered to ICU patients as a practice that hastens death, in addition to relieving unbearable suffering, and second, to assess ICU professionals' perceptions of assistance in dying. METHODS: A national survey with online questionnaires for ICU physicians and nursesaddressed through the French Society of Anesthesiology and Critical Care Medicine. RESULTS: A total of 956 ICU professionals responded to the survey (38% physicians and 62% nurses). Of these, 22% of physicians and 12% of nurses (p < 0.001) felt that the purpose of CDS was to hasten death. For 20% of physicians, CDS combined with terminal extubation was considered an assistance in dying. For 52% of ICU professionals, the current framework did not sufficiently cover the range of situations that occur in the ICU. A favorable opinion on the potential legalization of assistance in dying was observed in 83% of nurses and 71% of physicians (p < 0.001), with no preference between assisted suicide and euthanasia. CONCLUSION: Our findings highlight the tension between CDS and assisted suicide/euthanasia in the specific context of intensive care and suggest that ICU professionals would be supportive of a legislative evolution.

Predictive role of hepatic venous pressure gradient in bleeding events among patients with cirrhosis undergoing orthotopic liver transplantation.

Background & Aims: Major bleeding events during orthotopic liver transplantation (OLT) are associated with poor outcomes. The proportion of this risk related to portal hypertension is unclear. Hepatic venous pressure gradient (HVPG) is the gold standard for estimating portal hypertension. The aim of this study was to analyze the ability of HVPG to predict intraoperative major bleeding events during OLT in patients with cirrhosis. Methods: We retrospectively analyzed a prospective database including all patients with cirrhosis who underwent OLT between 2010 and 2020 and had liver and right heart catheterizations as part of their pre-transplant assessment. The primary endpoint was the occurrence of an intraoperative major bleeding event. Results: The 468 included patients had a median HVPG of 17 mmHg [interquartile range, 13-22] and a median MELD on the day of OLT of 16 [11-24]. Intraoperative red blood cell transfusion was required in 72% of the patients (median 2 units transfused), with a median blood loss of 1,000 ml [575-1,500]. Major intraoperative bleeding occurred in 156 patients (33%) and was associated with HVPG, preoperative hemoglobin level, severity of cirrhosis at the time of OLT (MELD score, ascites, encephalopathy), hemostasis impairment (thrombocytopenia, lower fibrinogen levels), and complications of cirrhosis (sepsis, acute-on-chronic liver failure). By multivariable regression analysis with backward elimination, HVPG, preoperative hemoglobin level, MELD score, and tranexamic acid infusion were associated with the primary endpoint. Three categories of patients were identified according to HVPG: low-risk (HVPG <16 mmHg), high-risk (HVGP ≥16 mmHg), and very high-risk (HVPG ≥20 mmHg). Conclusions: HVPG predicted major bleeding events in patients with cirrhosis undergoing OLT. Including HVPG as part of pre-transplant assessment might enable better anticipation of the intraoperative course. Impact and implications: Major bleeding events during orthotopic liver transplantation (OLT) are associated with poor outcomes but the proportion of this risk related to portal hypertension is unclear. Our work shows that hepatic venous pressure gradient (HVPG), the gold standard for estimating portal hypertension, is a strong predictor of major bleeding events and blood loss volume in patients with cirrhosis undergoing OLT. Three groups of patients can be identified according to their risk of major bleeding events: low-risk patients with HVPG <16 mmHg, high-risk patients with HVPG ≥16 mmHg, and very high-risk patients with HVPG ≥20 mmHg. HVPG could be systematically included in the pre-transplant assessment to anticipate intraoperative course and tailor patient management.

Team-family conflicts over end-of-life decisions in ICU: A survey of French physicians' beliefs.

INTRODUCTION: Conflicts between relatives and physicians may arise when decisions are being made about limiting life-sustaining therapies (LST). The aim of this study was to describe the motives for, and management of team-family conflicts surrounding LST limitation decisions in French adult ICUs. METHODS: Between June and October 2021, French ICU physicians were invited to answer a questionnaire. The development of the questionnaire followed a validated methodology with the collaboration of consultants in clinical ethics, a sociologist, a statistician and ICU clinicians. RESULTS: Among 186 physicians contacted, 160 (86%) answered all the questions. Conflicts over LST limitation decisions were mainly related to requests by relatives to continue treatments considered to be unreasonably obstinate by ICU physicians. The absence of advance directives, a lack of communication, a multitude of relatives, and religious or cultural issues were frequently mentioned as factors contributing to conflicts. Iterative interviews with relatives and proposal of psychological support were the most widely used tools in attempting to resolve conflict, while the intervention of a palliative care team, a local ethics resource or the hospital mediator were rarely solicited. In most cases, the decision was suspended at least temporarily. Possible consequences include stress and psychological exhaustion among caregivers. Improving communication and anticipation by knowing the patient's wishes would help avoid these conflicts. CONCLUSION: Team-family conflicts during LST limitation decisions are mainly related to requests from relatives to continue treatments deemed unreasonable by physicians. Reflection on the role of relatives in the decision-making process seems essential for the future.

Colonic perforation revealed by massive subcutaneous emphysema.

We present the case of an 82-year-old woman admitted in the intensive care unit with a septic shock caused by a liver abscess. She underwent an emergency laparotomy for abscess drainage and microbiological sampling. In the early postoperative period, she developed a massive subcutaneous emphysema (SE) extending from the abdomen to the head, without obvious cause. A surgical complication was suspected; thus, the patient underwent a second laparotomy which revealed a perforated peritonitis. SE is a rare presentation of perforated peritonitis, which should be known by critical care physicians in order to avoid a misdiagnosis of this life-threatening pathology.

The prognostic value of the neutrophil-to-lymphocyte ratio in critically ill cirrhotic patients.

BACKGROUND: Hospital death rates following ICU admission of cirrhotic patients remain high. Identifying patients at high risk of mortality after few days of aggressive management is imperative for providing adequate interventions. Herein, we aimed to evaluate the prognostic value of the neutrophil-to-lymphocyte ratio (NLR) combined with usual organ failure scores in the outcome prediction of cirrhotic patients hospitalized more than 3 days in ICU. METHODS: We conducted a retrospective bicentric study in two cohorts of cirrhotic patients hospitalized more than 3 days in French university hospital ICUs. At admission and day 3, we calculated several clinico-biological scores grading liver disease and organ failure severity and calculated the NLR. The primary outcome was 28-day mortality. RESULTS: The test cohort included 116 patients. At day 28, 43 (37.1%) patients had died. Variations of MELD score (ΔMELD), SOFA score (ΔSOFA), CLIF-SOFA score (ΔCLIF-SOFA) and NLR (ΔNRL) between admission and day 3 were significantly associated with 28-day mortality in univariate analysis. When included in bivariate analysis ΔNLR remained a significant predictor of 28-day mortality independently of these severity scores. Kaplan-Meier curves and statistics using reclassification methods showed a better 28-day mortality risk prediction using ΔNRL in association with ΔSOFA in comparison to ΔSOFA alone. These results were confirmed in an external validation cohort, including 101 critically ill cirrhotic patients. CONCLUSIONS: ΔNLR is an independent predictor of mortality in the critically ill cirrhotic patients' population who requires intensive care supportive treatment and should be used in association with ΔSOFA as a prognostic biomarker.

Risk Factors and Impact of Perioperative Prophylaxis on the Risk of Extended-spectrum ß-Lactamase-producing Enterobacteriaceae-related Infection Among Carriers Following Liver Transplantation.

BACKGROUND: Extended-spectrum ß-lactamase-producing Enterobacteriaceae (ESBL-E) carriage is frequent among liver transplant (LT) recipients, thereby fostering a large empirical carbapenem prescription. However, ESBL-E infections occur in only 10%-25% of critically ill patients with rectal colonization. Our aim was to identify risk factors for post-LT ESBL-E infection in colonized patients. The effect of perioperative antimicrobial prophylaxis (AP) was also analyzed in patients with prophylaxis lasting <48 hours and without proven intraoperative infection. METHODS: Retrospective study from a prospective database including patients with a positive ESBL-E rectal screening transplanted between 2010 and 2016. RESULTS: Among the 749 patients transplanted, 100 (13.3%) were colonized with an ESBL-E strain. Thirty-nine (39%) patients developed an infection related to the same ESBL-E (10 pulmonary, 11 surgical site, 13 urinary, 5 bloodstream) within 11 postoperative days in median. Klebsiella pneumoniae carriage, model for end-stage liver disease ≥25, preoperative spontaneous bacterial peritonitis prophylaxis, and antimicrobial exposure during the previous month were independent predictors of ESBL-E infection. We propose a colonization to infection risk score built on these variables. The prevalence of infection for colonization to infection score of 0, 1, 2, and ≥3 were 7.4%, 26.3%, 61.9%, and 91.3%, respectively. Of note, the incidence of post-LT ESBL-E infection was lower in case of perioperative AP targeting colonizing ESBL-E (P = 0.04). CONCLUSIONS: Thirty-nine percentage of ESBL-E carriers develop a related infection after LT. We identified predictors for ESBL-E infection in carriers that may help in rationalizing carbapenem prescription. Perioperative AP targeting colonizing ESBL-E may be associated with a reduced risk of post-LT ESBL-E infections.

Development of the Gastrointestinal Dysfunction Score (GIDS) for critically ill patients - A prospective multicenter observational study (iSOFA study).

BACKGROUND & AIMS: To develop a five grade score (0-4 points) for the assessment of gastrointestinal (GI) dysfunction in adult critically ill patients. METHODS: This prospective multicenter observational study enrolled consecutive adult patients admitted to 11 intensive care units in nine countries. At all sites, daily clinical data with emphasis on GI clinical symptoms were collected and intra-abdominal pressure measured. In five out of 11 sites, the biomarkers citrulline and intestinal fatty acid-binding protein (I-FABP) were measured additionally. Cox models with time-dependent scores were used to analyze associations with 28- and 90-day mortality. The models were estimated with stratification for study center. RESULTS: We included 540 patients (224 with biomarker measurements) with median age of 65 years (range 18-94), the Simplified Acute Physiology Score II score of 38 (interquartile range 26-53) points, and Sequential Organ Failure Assessment (SOFA) score of 6 (interquartile range 3-9) points at admission. Median ICU length of stay was 3 (interquartile range 1-6) days and 90-day mortality 18.9%. A new five grade Gastrointestinal Dysfunction Score (GIDS) was developed based on the rationale of the previously developed Acute GI Injury (AGI) grading. Citrulline and I-FABP did not prove their potential for scoring of GI dysfunction in critically ill. GIDS was independently associated with 28- and 90-day mortality when added to SOFA total score (HR 1.40; 95%CI 1.07-1.84 and HR 1.40; 95%CI 1.02-1.79, respectively) or to a model containing all SOFA subscores (HR 1.48; 95%CI 1.13-1.92 and HR 1.47; 95%CI 1.15-1.87, respectively), improving predictive power of SOFA score in all analyses. CONCLUSIONS: The newly developed GIDS is additive to SOFA score in prediction of 28- and 90-day mortality. The clinical usefulness of this score should be validated prospectively. TRIAL REGISTRATION: NCT02613000, retrospectively registered 24 November 2015.

Non-readmission decisions in the intensive care unit under French rules: A nationwide survey of practices.

PURPOSE: We investigated, using a multicentre survey of practices in France, the practices of ICU physicians concerning the decision not to readmit to the ICU, in light of current legislation. MATERIALS AND METHODS: Multicentre survey of practices among French ICU physicians via electronic questionnaire in January 2016. Questions related to respondents' practices regarding re-admission of patients to the ICU and how these decisions were made. Criteria were evaluated by the health care professionals as regards importance for non-readmission. RESULTS: In total, 167 physicians agreed to participate, of whom 165 (99%) actually returned a completed questionnaire from 58 ICUs. Forty-five percent were aged <35 years, 74% were full-time physicians. The findings show that decisions for non-readmission are taken at the end of the patient's stay (87%), using a collegial decision-making procedure (89% of cases); 93% reported that this decision was noted in the patient's medical file. While 73% indicated that the family/relatives were informed of non-readmission decisions, only 29% reported informing the patient, and 91% considered that non-readmission decisions are an integral part of the French legislative framework. CONCLUSION: This study shows that decisions not to re-admit a patient to the ICU need to be formally materialized, and anticipated by involving the patient and family in the discussions, as well as the other healthcare providers that usually care for the patient. The optimal time to undertake these conversations is likely best decided on a case-by-case basis according to each patient's individual characteristics.

Ethical aspects of admission or non-admission to the intensive care unit.

The question of admission and non-admission to the intensive care unit (ICU) raises several ethical questions. There is a fine line between the risk of loss-of-opportunity for the patient in case of non-admission, and the risk of unreasonable therapeutic obstinacy, in case of unjustified admission. Similar difficulties arise in decisions regarding re-admission or non-re-admission, with the sole difference that the intensivists already know the patient and his/her medical history. This information can help inform the decision when re-admission is being considered. Intensive, i.e., life-sustaining care should be implemented after shared reflection involving the caregivers, the patient and the family, and the same applies for non-implementation of these same therapies. Anticipating admission or non-admission to the ICU in case of acute organ failure, or in case of potential deterioration represents a major challenge for our discipline in the coming years.

Prognosis of cirrhotic patients admitted to the general ICU.

BACKGROUND: The prognosis of cirrhotic patients admitted to the ICU is considered to be poor but has been mainly reported in liver ICU. We aimed to describe the prognosis of cirrhotic patients admitted to a general ICU, to assess the predictors of mortality in this population, and, finally, to identify a subgroup of patients in whom intensive care escalation might be discussed. RESULTS: We performed a retrospective monocentric study of all cirrhotic patients consecutively admitted between 2002 and 2014 in a general ICU in a regional university hospital. Two hundred and eighteen cirrhotic patients were admitted to the ICU. The 28-day and 6-month mortality rates were 53 and 74 %, respectively. Among the 115 patients who were discharged from ICU, only eight patients underwent liver transplantation, whereas 48 had no clear contraindication. Multivariable analyses on 28-day mortality identified three independent variables, incorporated into a new three-variable prognostic model as follows: SOFA ≥ 12 (OR 4.2 [2.2-8.0]; 2 points), INR ≥ 2.6 (OR 2.5 [1.3-4.8]; 1 point), and renal replacement therapy (OR 2.3 [1.1-5.1]; 1 point). For a value of the score at 4 (16 % of patients), 28-day and 3-month mortality rates were 91 and 100 %, respectively. An external validation of the score among 149 critically ill cirrhotic patients showed a good accuracy for predicting in-ICU mortality. CONCLUSIONS: Mortality of cirrhotic patients admitted to a general ICU was comparable to that of other studies. A pragmatic score integrating the SOFA score, INR, and the need for extrarenal epuration was strongly associated with mortality. Among the 16 % of patients presenting with score 4 at ICU admission, 100 % died in the 3-month follow-up period. The prognostic evaluation on day 3 remains essential for the majority of patients. However, this score calculable at ICU admission might identify patients in whom the benefit of intensive care escalation should be discussed, in particular when liver transplantation is contraindicated.

[Successful treatment of extreme hypernatremia by continuous veno-venous hemodiafiltration]. / Une hypernatrémie extrême traitée avec succès par hémodiafiltration veino-veineuse continue.

Extreme hypernatremia in intensive care unit are frequently associated with a poor prognosis and their treatment, when associated with acute renal failure, is not consensual. We report the case of a 39-year-old man admitted in our intensive care unit for coma who presented extreme hyperosmolar hypernatremia (sodium 180 mmol/L, osmolarity 507 mOsm/L) associated with acute renal failure (urea 139.3 mmol/L, creatinine 748 µmol/L) and many other metabolic abnormalities. He was treated with hypotonic fluid administration and continuous renal replacement therapy (veno-venous hemodiafiltration) using an industrial dialysate fluid. Natremia was controlled by modulating intravenous water and sodium intake according to biological data. After 10 days, continuous renal replacement therapy was stopped and neurological exam was normal. Continuous veno-venous hemodiafiltration may be useful for treatment of extreme hypernatremia by allowing gradual correction of fluid and electrolyte disorders.