BackgroundIn neurocritical care units (NCCU), decisions to withhold life-sustaining therapies are sometimes influenced by anticipated disability and quality of life (QoL) impairment, particularly when further interventions are deemed futile. However, health-related QoL (HRQoL) is inherently subjective and does not always correlate with disability levels. This qualitative, noninterventional study aimed to assess the relevance of HRQoL in ethical decision-making using interpretative phenomenological analysis (IPA) to investigate subjective HRQoL.MethodsPatients were interviewed by a single intensivist to assess their subjective QoL 2 years after their stay in an NCCU following an acute brain injury (ABI). The intensivist directed the interview toward HRQoL using a guide comprising limited and mostly open-ended questions. Audio recordings of the interviews were transcribed verbatim into Word narratives, analyzed in depth by two intensivists using IPA methodology and NVivo 14 software to enable exploration of patients' lived experiences and personal QoL assessments.ResultsA total of 14 patients were invited to the follow-up appointment, 7 for whom life-sustaining treatment had been withheld in the NCCU and 7 matched patients for whom this decision had not been made. Among the nine patients finally included, life-sustaining treatment had been withheld in four cases. Patients varied greatly in how they perceived and valued their QoL. While most valued relationships and independence, they expressed these values in different ways. Frustration with disability and support from relatives emerged as key motivators for rehabilitation. Despite their challenges, patients expressed gratitude for survival and pride in their progress and daily achievements. Overall, their experiences highlighted the deeply personal and subjective nature of disability and QoL assessment.ConclusionsHRQoL after ABI is highly subjective and should be considered with great caution in decisions to withhold life-sustaining treatment in NCCU. Further studies are warranted to improve outcome assessment after ABI and aid ethical decision-making.

Severe traumatic brain injury (sTBI) is a major cause of mortality and long-term disability, and early prognostic evaluation is essential to support therapeutic decision-making in intensive care. MRI is the reference imaging modality for detecting diffuse axonal injury (DAI), yet the prognostic accuracy and reproducibility of MRI-based radiologic scoring systems remain uncertain. This study aimed to compare the prognostic performance of five published MRI-based DAI scores for predicting 1-year neurological outcome, in a cohort of sTBI patients admitted to intensive care. We analyzed adult patients with sTBI included in the prospective MRI-COMA cohort (NCT00577954). Inclusion criteria were age ≥ 18 years, admission to ICU for sTBI, and absence of command following within 7 days after sedation withdrawal. Brain MRI was performed between day 7 and day 35 post-injury using standardized T1, FLAIR, T2*/SWI, and diffusion-weighted sequences. Three blinded evaluators (one neurointensivist, two neuroradiologists) independently applied five radiologic scores (Adams, Firsching, Hamdeh, Stockholm, Trondheim). The primary endpoint was the ability of each MRI-based radiologic score to predict 1-year neurological outcome, dichotomized as favorable (GOSE 5–8) or unfavorable (GOSE 1–4). Inter-rater reliability was quantified using Cohen’s and Fleiss’ kappa coefficients. Between 2007 and 2023, 443 patients were screened and 185 met eligibility criteria. At one year, 111 patients (59

A reliable outcome prognostication tool for patients in coma of various etiologies would facilitate ICU treatment by providing objective information to caregivers and patients' relatives. This study aimed to predict outcome based on supervised machine learning and magnetic resonance diffusion tensor imaging (DTI) metrics. In this multicenter international study, a training set of 531 patients not responding to simple orders at day 5 after coma onset underwent diffusion-weighted MRI between day 5 and 45. A classifier was developed using DTI metrics, patient age, and delay between admission and MRI as features. Unfavorable outcome (UFO) was defined as GOSE 1–4 at one year. Three prognosis areas were defined: a “red” zone (specificity for UFO above 95

In the growing field of neuroprognosis in intensive care patients, the use of diffusion tensor imaging (DTI) MRI to quantify white matter integrity is particularly promising, with solid demonstrations for predicting outcomes in cardiac arrest and severe traumatic brain injury. However, as with any new technique available to practitioners in routine practice, it is essential to provide clinicians and future users with the validity criteria they need to master for an appropriate interpretation of their results. This pragmatic review is aimed at providing clinicians and future users with a comprehensive understanding of the medical and technical conditions necessary for the accurate interpretation of DTI-derived metrics. Although DTI offers significant potential, its reliability can be influenced by various factors, including previous neurological disorders, acquisition artifacts, and data preprocessing failures. This review highlights the importance of considering classical MRI contraindications and the impact of physiological conditions such as age and the timing of MRI scans. Additionally, it discusses the influence of preexisting neurological disorders and associated lesions on DTI metrics, which can lead to classification biases and prognostication errors. Technical considerations, including the necessity for rigorous quality control and preprocessing steps, are also emphasized to ensure the precision and accuracy of DTI metrics. By addressing these factors, this review is aimed at enhancing the interpretability and clinical utility of DTI in neuroprognostication, ultimately facilitating more informed decision-making for intensive care unit patients. Furthermore, ongoing research and data sharing are needed to refine DTI techniques and improve their predictive accuracy in various clinical contexts.

In the growing field of neuroprognosis in intensive care patients, the use of diffusion tensor imaging (DTI) MRI to quantify white matter integrity is particularly promising, with solid demonstrations for predicting outcomes in cardiac arrest and severe traumatic brain injury. However, as with any new technique available to practitioners in routine practice, it is essential to provide clinicians and future users with the validity criteria they need to master for an appropriate interpretation of their results. This pragmatic review is aimed at providing clinicians and future users with a comprehensive understanding of the medical and technical conditions necessary for the accurate interpretation of DTI-derived metrics. Although DTI offers significant potential, its reliability can be influenced by various factors, including previous neurological disorders, acquisition artifacts, and data preprocessing failures. This review highlights the importance of considering classical MRI contraindications and the impact of physiological conditions such as age and the timing of MRI scans. Additionally, it discusses the influence of preexisting neurological disorders and associated lesions on DTI metrics, which can lead to classification biases and prognostication errors. Technical considerations, including the necessity for rigorous quality control and preprocessing steps, are also emphasized to ensure the precision and accuracy of DTI metrics. By addressing these factors, this review is aimed at enhancing the interpretability and clinical utility of DTI in neuroprognostication, ultimately facilitating more informed decision-making for intensive care unit patients. Furthermore, ongoing research and data sharing are needed to refine DTI techniques and improve their predictive accuracy in various clinical contexts.

Background: Core body temperature (CBT) plays a pivotal role in determining the prognosis of patients with neurological impairments. This study aimed to develop and evaluate a machine learning (ML) algorithm capable of forecasting CBT, as well as predicting fever and hypothermia, several hours in advance. Methods: We conducted a multicenter retrospective observational study in three mixed intensive care units (ICUs): two university hospitals in France (Hopital de la Timone and Nord), and one teaching military hospital in France (Sainte-Anne). We evaluated several prediction methods, including Neural Networks, Gradient Boosting, Random Forests, linear regression models, LSTM, and XGBoost. Inputs consisted of past temperature measurements, blood pressure, heart rate, and time of day. Results: Data from 10,189 ICU patients were analyzed. A training cohort (n = 5,146) from two ICUs was used to develop the models, and an independent evaluation cohort (n = 5,043) from the third ICU was used for testing. XGBoost consistently demonstrated the highest predictive performance for both fever and hypothermia. Sensitivity for fever (and hypothermia) prediction was 93.2 % (90.2 %) at 1 h, 86.6 % (82.3 %) at 2 h, and 76.8 % (70.3 %) at 4 h. Prediction of CBT values yielded Root Mean Square Errors of 0.19 degrees C, 0.31 degrees C, and 0.46 degrees C at 1, 2, and 4 h, respectively. Conclusion: This is the first large multicenter study to evaluate the contribution of ML to CBT prediction in ICU patients. Our findings show that fever and hypothermia can be reliably detected up to four hours before their occurrence, paving the way for more proactive and personalized patient management.

Delayed graft function is the most frequent early complication of kidney transplantation. Pretreatment of kidney donors with cyclosporine has decreased delayed graft function in animal studies by reducing ischemia–reperfusion graft injuries. No randomized clinical trials have assessed the efficacy of cyclosporine pretreatment of brain-dead donors in reducing delayed graft function. In this multicenter randomized, double-blind, and placebo-controlled trial, brain-dead donors were randomized (1:1) to receive either 2.5 mg/kg of cyclosporine or a glucose placebo infusion. The kidney transplant candidates were allocated through their donor assignment. The primary outcome was the occurrence of delayed graft function (DGF), defined as the need for at least one hemodialysis within the 7 days after kidney transplantation. Secondary outcomes included early graft function parameters within the 7 days post-transplantation, and 1-year graft and recipient survival. Between December 17, 2017 and March 3, 2023, 258 donors/331 recipients in the placebo group and 238 donors/312 recipients in the cyclosporine group were included in the modified intention-to-treat analysis. DGF occurred in 46 recipients (13.9

OBJECTIVE:This study aimed to identify predictors of epidural blood patch (EBP) success in patients with intracranial hypotension syndrome. BACKGROUND:The epidural blood patch remains the gold standard treatment for intracranial hypotension syndrome, yet its effectiveness varies, and predictors of sustained success remain uncertain. METHODS:We performed a single-center retrospective cohort study. We analyzed 139 epidural blood patches performed for non-obstetric intracranial hypotension syndrome between April 2015 and July 2025. Demographic, clinical, biological, radiological, and procedural data were collected. Complete symptom resolution or highly significant improvement at 1 month was defined as treatment success. RESULTS:Among 93 patients, 1-month complete success was achieved after 31% of procedures (43/139). Early improvement within 48 h was associated with higher odds of 1-month EBP success (adjusted odds ratio [aOR] = 9.70, 95% confidence interval [CI]: 3.66-28.96; p < 0.001) as was the occurrence of rebound headache (aOR = 5.34, 95% CI: 1.27-26.44; p = 0.028). Conversely, symptom exacerbation during the Valsalva maneuver was associated with lower odds of 1-month EBP success (aOR = 0.30, 95% CI: 0.10-0.81; p = 0.020). Lower baseline platelet and fibrinogen levels and osteophytic leaks were negatively associated with early effectiveness but not with 1-month outcome. Targeted epidural blood patches and shorter delay from symptom onset to procedure were associated with higher success rates in univariate analysis, though not independently. Model performance was robust with an area under the curve (AUC) of 0.88 (95% CI: 0.80-0.90). No infectious complications were observed; rebound headaches and transient back pain were the most common secondary events. CONCLUSION:Early clinical improvement and rebound headache are strong positive predictors of durable epidural blood patch effectiveness in intracranial hypotension syndrome, whereas Valsalva-related symptom worsening indicates a higher risk of failure. Radiological severity did not predict outcomes in our study. These easily identifiable clinical factors can assist in individualized management and reduce unnecessary repeat procedures.

Cefotaxime is a widely prescribed cephalosporin antibiotic used to treat various infections. It is mainly eliminated unchanged by the kidney through tubular secretion and glomerular filtration. Therefore, a reduction of kidney function may increase exposure to the drug and induce toxic side effects. The objectives of this study were to develop a physiologically based pharmacokinetic (PBPK) model of cefotaxime in healthy European adults, to mechanistically describe the impact of chronic kidney disease (CKD) on cefotaxime pharmacokinetics, and to assess the applicability of the model to patients requiring intensive care. Using PK-Sim® software, we developed a PBPK model for cefotaxime, including basolateral and apical renal transporters and renal esterases, in healthy subjects and then extrapolated to patients with CKD by incorporating pathophysiological changes and reductions in activity of drug-metabolizing enzymes and transporters into the model. We then evaluated the predictive performance of the model in patients requiring intensive care using clinical routine data. Model predictions were considered adequate in healthy subjects and patients with CKD, with predicted-to-observed area under the curve ratios within the two-fold acceptance criterion. Mean prediction error and mean absolute prediction error did not exceed ± 30 and 30

Chaque année, le Conseil national des universités (CNU Santé) organise, sur instruction du ministère de l’Enseignement supérieur et de la Recherche, les concours de mise sur liste d’aptitude aux fonctions de maîtres de conférences et professeurs des universités – praticiens hospitaliers (MCU-PH, PU-PH). En 2025, la sous-section 48-01 « Anesthésiologie-Réanimation et Médecine Périopératoire » (ARMPO) a auditionné quinze candidats (3 MCU-PH, 12 PU-PH). Le concours est précédé d’une visite sur site par deux rapporteurs désignés, qui évaluent l’intégration au sein de l’établissement, les aptitudes scientifiques, pédagogiques et dans la conduite d’équipe du candidat. Le jour du concours, l’audition se déroule en deux étapes : présentation du parcours et du projet hospitalo-universitaire du candidat, puis échange avec le jury du CNU. Une épreuve de pédagogie (leçon de 20minutes) et la rédaction préalable d’une station d’Examen clinique objectif structuré (ECOS) complètent l’évaluation. La promotion 2025 illustre la diversité et la richesse des profils : anesthésie spécialisée, réanimation chirurgicale et polyvalente, innovations pédagogiques, recherche translationnelle, intelligence artificielle, éthique clinique, organisation des soins et développement durable. Les projets couvrent des thématiques variées : optimisation hémodynamique, médecine personnalisée cardiovasculaire, prévention des dysfonctions d’organes, soins périopératoires intégrés, simulation haute-fidélité, neurosciences, allergo-anesthésie, traumatologie, transplantation, ischémie-reperfusion pulmonaire, immunologie du sepsis. À travers la présentation de ces parcours, la sous-section 48-01 témoigne de la vitalité scientifique, pédagogique et hospitalière de l’ARMPO en France, discipline à l’interface de multiples expertises et résolument tournée vers l’innovation, la collaboration interdisciplinaire et la formation des générations futures.

Importance:Whether the use of inhaled or intravenous sedation affects outcomes differentially in mechanically ventilated adults with acute respiratory distress syndrome (ARDS) is unknown. Objective:To determine the efficacy and safety of inhaled sevoflurane compared with intravenous propofol for sedation in patients with ARDS. Design, Setting, and Participants:Phase 3 randomized, open-label, assessor-blinded clinical trial conducted from May 2020 to October 2023 with 90-day follow-up. Adults with early moderate to severe ARDS (defined by a ratio of Pao2 to the fraction of inspired oxygen of <150 mm Hg with a positive end-expiratory pressure of ≥8 cm H2O) were enrolled in 37 French intensive care units. Interventions:Patients were randomized to a strategy of inhaled sedation with sevoflurane (intervention group) or to a strategy of intravenous sedation with propofol (control group) for up to 7 days. Main Outcomes and Measures:The primary end point was the number of ventilator-free days at 28 days; the key secondary end point was 90-day survival. Results:Of 687 patients enrolled (mean [SD] age, 65 [12] years; 30% female), 346 were randomized to sevoflurane and 341 to propofol. The median total duration of sedation was 7 days (IQR, 4 to 7) in both groups. The number of ventilator-free days through day 28 was 0.0 days (IQR, 0.0 to 11.9) in the sevoflurane group and 0.0 days (IQR, 0.0 to 18.7) in the propofol group (median difference, -2.1 [95% CI, -3.6 to -0.7]; standardized hazard ratio, 0.76 [95% CI, 0.50 to 0.97]). The 90-day survival rates were 47.1% and 55.7% in the sevoflurane and propofol groups, respectively (hazard ratio, 1.31 [95% CI, 1.05 to 1.62]). Among 4 secondary outcomes, sevoflurane was associated with higher 7-day mortality (19.4% vs 13.5%, respectively; relative risk, 1.44 [95% CI, 1.02 to 2.03]) and fewer intensive care unit-free days through day 28 (median, 0.0 [IQR, 0.0 to 6.0] vs 0.0 [IQR, 0.0 to 15.0]; median difference, -2.5 [95% CI, -3.7 to -1.4]) compared with propofol. Conclusions and Relevance:Among patients with moderate to severe ARDS, inhaled sedation with sevoflurane resulted in fewer ventilator-free days at day 28 and lower 90-day survival than sedation with propofol. Trial Registration:ClinicalTrials.gov Identifier: NCT04235608.

Whether the use of inhaled or intravenous sedation affects outcomes differentially in mechanically ventilated adults with acute respiratory distress syndrome (ARDS) is unknown. To determine the efficacy and safety of inhaled sevoflurane compared with intravenous propofol for sedation in patients with ARDS. Phase 3 randomized, open-label, assessor-blinded clinical trial conducted from May 2020 to October 2023 with 90-day follow-up. Adults with early moderate to severe ARDS (defined by a ratio of Pao2 to the fraction of inspired oxygen of <150 mm Hg with a positive end-expiratory pressure of ≥8 cm H2O) were enrolled in 37 French intensive care units. Patients were randomized to a strategy of inhaled sedation with sevoflurane (intervention group) or to a strategy of intravenous sedation with propofol (control group) for up to 7 days. The primary end point was the number of ventilator-free days at 28 days; the key secondary end point was 90-day survival. Of 687 patients enrolled (mean [SD] age, 65 [12] years; 30% female), 346 were randomized to sevoflurane and 341 to propofol. The median total duration of sedation was 7 days (IQR, 4 to 7) in both groups. The number of ventilator-free days through day 28 was 0.0 days (IQR, 0.0 to 11.9) in the sevoflurane group and 0.0 days (IQR, 0.0 to 18.7) in the propofol group (median difference, −2.1 [95% CI, −3.6 to −0.7]; standardized hazard ratio, 0.76 [95% CI, 0.50 to 0.97]). The 90-day survival rates were 47.1% and 55.7% in the sevoflurane and propofol groups, respectively (hazard ratio, 1.31 [95% CI, 1.05 to 1.62]). Among 4 secondary outcomes, sevoflurane was associated with higher 7-day mortality (19.4% vs 13.5%, respectively; relative risk, 1.44 [95% CI, 1.02 to 2.03]) and fewer intensive care unit–free days through day 28 (median, 0.0 [IQR, 0.0 to 6.0] vs 0.0 [IQR, 0.0 to 15.0]; median difference, –2.5 [95% CI, –3.7 to –1.4]) compared with propofol. Among patients with moderate to severe ARDS, inhaled sedation with sevoflurane resulted in fewer ventilator-free days at day 28 and lower 90-day survival than sedation with propofol. ClinicalTrials.gov Identifier: NCT04235608

β-lactam-induced neurotoxicity in critical care patients can compromise clinical outcomes. Despite the growing use of therapeutic drug monitoring (TDM) for β-lactams, clear toxicity thresholds remain undefined, leaving clinicians uncertain about dosing adjustments when adverse effects occur. Identifying a relevant and easily detectable neurophysiological biomarker for β-lactam exposure would improve monitoring and prevent serious complications. In a prospective multicenter, non-interventional study, we analyzed electroencephalographic (EEG) signals of 56 patients hospitalized in intensive care units (ICUs) receiving continuous infusions of five β-lactams (meropenem, piperacillin/tazobactam, cefepime, cefotaxime, or ceftazidime). We applied a time frequency decomposition on these EEG data to investigate quantitatively the power of neural dynamics across frequencies ranging from 1 to 45 Hz. We used a multivariate pattern decoding method to correlate the β-lactam exposure and Sepsis-related Organ Failure Assessment (SOFA) scores with the neural activity. β-lactam exposure correlated with increased β-low γ neural dynamics (20–40 Hz) (p < 0.001, FDR corrected), independent of other clinical factors or medications. β-neural activity was most pronounced in central electrodes (C3: r = 0.20, p < 0.01; C4: r = 0.26, p < 0.01) and the right frontal electrode (Fp2: r = 0.12, p = 0.02). Lower θ-α activity (3.5–5 Hz and 12–18 Hz) was associated with higher SOFA scores (p < 0.001, FDR corrected). No significant correlations were observed between other drugs (opioids, anti-seizure medications, and psychotropics) and β or θ-α dynamics. These results suggest that β neural dynamics represent a potential biomarker for β-lactam exposure in ICU patients. They highlight the potential of quantitative EEG and advanced multivariate decoding methods to identify subtle neurophysiological features that are otherwise difficult to detect. ClinicalTrials.gov ID NCT03339869. Registered 14 September 2017.

The National Council of Universities (CNU) for Health, under the Ministry of Higher Education and Research, annually conducts competitive examinations to appoint associate professors (MCU-PH) and full professors (PU-PH) in medicine. In 2025, subsection 48-01 "Anesthesiology, Critical Care, and Perioperative Medicine" (ACCPM) evaluated 15 candidates (3 MCU-PH, 12 PU-PH) through a rigorous process. The selection includes an on-site visit by two appointed reviewers-assessing integration within the institution, scientific and educational contributions, and leadership, followed by a two-stage audition: presentation of the candidate's career and academic project, and an interactive discussion with the CNU panel. On the day of the examination, candidates also deliver a 20-minute teaching session and submit a previously prepared Objective Structured Clinical Examination (OSCE) station. The 2025 cohort reflects the breadth of ACCPM expertise: specialized anesthesia, surgical and general intensive care, perioperative medicine, and multidisciplinary leadership. Research themes span hemodynamic optimization, personalized cardiovascular medicine, prevention of organ dysfunction, translational and clinical neurosciences, high-fidelity simulation, allergo-anesthesia, trauma care, transplantation, pulmonary ischemia-reperfusion, sepsis immunology, and applications of artificial intelligence in critical care. These profiles also demonstrate a commitment to innovation in medical education, ethical practice, sustainable healthcare, and integrated care pathways. Collectively, they highlight the scientific vitality and strategic vision of ACCPM in France, an inherently interdisciplinary field dedicated to improving patient outcomes, advancing research, and training the next generation of perioperative and critical care specialists.

The fibroproliferative stage and persistent inflammation of acute respiratory distress syndrome (ARDS) are key factors leading to either the resolution of the syndrome or fibrosis. Previous studies suggest that a corticosteroid therapy promotes the evolution of ARDS toward an adapted repair process whereas others suggest that this therapy increases the risk of death if it starts more than 14 days after ARDS onset. Since the efficacy and safety of delayed 2 mg/kg methylprednisolone therapy in patients with ARDS is a matter of debate, we performed this observational multicentric retrospective study. We analysed the data of 392 patients with ARDS who received 2 mg/kg methylprednisolone therapy. The primary endpoint was mortality six months after 2 mg/kg methylprednisolone therapy was started. The secondary endpoints included mortality 60 days after the corticosteroid therapy initiation and the number of ventilator-free days (VFDs) and intensive care unit (ICU)-free days. We investigated the occurrence of complications such as ventilator-acquired pneumonia (VAP), septic shock and gastrointestinal bleeding arising after the start of the protocol. A total of 189 (48.2%) patients received 2 mg/kg methylprednisolone therapy within the first 14 days of ARDS onset. A total of 203 (51.8%) patients received it more than 14 days included post-ARDS-onset. The mortality rate six months after the initiation of 2 mg/kg methylprednisolone therapy was 51.9% in the early initiation group and 52.2% in the late initiation group (p = 0.942). The mortality rate 60 days after the initiation of 2 mg/kg methylprednisolone therapy was 47.1% in the early group and 47.3% in the late group (p = 0.968). There was no significant difference in the number of VFDs (p = 0.336) or ICU-free days (p = 0.175) 60 days after the start of the 2 mg/kg protocol. Initiating the protocol 14 days after the onset of ARDS seemed to be associated with more complications (p < 0.001). Late initiation was associated with greater occurrence of VAP (p = 0.018) or gastrointestinal bleeding (p = 0.012). These results suggest that an initiation of 2 mg/kg methylprednisolone therapy after 14 days from ARDS onset is not associated with an increased risk of death as compared with initiation prior to day 14. Delayed 2 mg/kg methylprednisolone therapy in patients with persistent ARDS should be considered.

External ventricular drain (EVD)-associated infections are a significant cause of morbidity and mortality. Nosocomial meningitis (NM) poses diagnostic challenges, and its prognosis heavily relies on the timely initiation of treatment. The aim of this study was to investigate the epidemiology of NM and risk factors in ICU patients. We conducted a retrospective single-center cohort study of adult patients who received an EVD in a French ICU between 2018 and 2022. Patients were classified into those with NM or without meningitis based on biological and microbiological criteria. We assessed risk factors related to the patient, the device, and the primary pathology, treatment regimens, length of stay, and survival. Of 275 patients with EVD, 32 (11.6

BACKGROUND:Proximal femur fractures (PFF) are common in the elderly, representing a significant public health issue. This study aims to define the epidemiology and morbidity of PFF and identify factors associated with 90-day mortality in patients with osteoporotic PFF. METHODS:We conducted a retrospective, bicentric, observational study in Marseille from November 2018 to June 2023, including patients operated for osteoporotic PFF. Clinical, biological, therapeutic, and socio-economic data were collected to analyse their influence on 90-day mortality and construct a mortality predictive model using a neural network. RESULTS:During the study period, 2442 patients were included, the mean age at diagnosis being 81 (13.7) years. The 279 (11.4%)non-survivors at 90 days were older (87.2 (0.8) vs. 79.8 (0.3) years; p < 0.0001), predominantly males (36.6% vs. 27.8%; p = 0.003), and had higher Charlson score (1[0-3] vs. 1[0-2]; p < 0.0001) and ASA score (3[3-3] vs. 3[2-3]; p < 0.0001). No significant differences were found in the use of cement, type of anaesthesia, and socio-economic level. A neural network predictive model for 90-day mortality included age, gender, ASA score, perioperative confusion, and haemoglobin, creatinine, and albumin at inclusion. The model performed with an area under the receiving operating characteristic curve of 0.91 [0.91-0.92], sensitivity of 50.5%, specificity of 75.0%, positive predictive value of 20.6%, and negative predictive value of 92.1% on the test set. CONCLUSION:Our results provide interesting elements to optimize the perioperative management of these patients. The future perspectives include validating the predictive model on external cohorts and integrating new variables to improve its accuracy.