Acute encephalopathy in the ICU: a practical approach.

PURPOSE OF REVIEW: Acute encephalopathy (AE) - which frequently develops in critically ill patients with and without primary brain injury - is defined as an acute process that evolves rapidly and leads to changes in baseline cognitive status, ranging from delirium to coma. The diagnosis, monitoring, and management of AE is challenging. Here, we discuss advances in definitions, diagnostic approaches, therapeutic options, and implications to outcomes of the clinical spectrum of AE in ICU patients without primary brain injury. RECENT FINDINGS: Understanding and definitions of delirium and coma have evolved. Delirium is a neurocognitive disorder involving impairment of attention and cognition, usually fluctuating, and developing over hours to days. Coma is a state of unresponsiveness, with absence of command following, intelligible speech, or visual pursuit, with no imaging or neurophysiological evidence of cognitive motor dissociation. The CAM-ICU(-7) and the ICDSC are validated, guideline-recommended tools for clinical delirium assessment, with identification of clinical subtypes and stratification of severity. In comatose patients, the roles of continuous EEG monitoring and neuroimaging have grown for the early detection of secondary brain injury and treatment of reversible causes. SUMMARY: Evidence-based pharmacologic treatments for delirium are limited. Dexmedetomidine is effective for mechanically ventilated patients with delirium, while haloperidol has minimal effect of delirium but may have other benefits. Specific treatments for coma in nonprimary brain injury are still lacking.

Antimicrobial consumption and drug utilization patterns among COVID-19 and non-COVID-19 patients.

OBJECTIVES: To understand differences in antimicrobial use between COVID-19 and non-COVID-19 patients. To compare two metrics commonly used for antimicrobial use: Defined Daily Dose (DDD) and Days of Therapy (DOT). To analyse the order in which antimicrobials were prescribed to COVID-19 patients using process mining techniques. METHODS: We analysed data regarding all ICU admissions from 1 January 2018 to 14 September 2020, in 17 Brazilian hospitals. Our main outcome was the antimicrobial use estimated by the DDD and DOT (Days of Therapy). We compared clinical characteristics and antimicrobial consumption between COVID-19 and non-COVID-19 patients. We used process mining to evaluate the order in which the antimicrobial schemes were prescribed to each COVID-19 patient. RESULTS: We analysed 68 405 patients admitted before the pandemic, 12 319 non-COVID-19 patients and 3240 COVID-19 patients. Comparing those admitted during the pandemic, the COVID-19 patients required advanced respiratory support more often (42% versus 12%). They also had longer ICU length of stay (6 versus 3 days), higher ICU mortality (18% versus 5.4%) and greater use of antimicrobials (70% versus 39%). Most of the COVID-19 treatments started with penicillins with ß-lactamase inhibitors (30%), third-generation cephalosporins (22%), or macrolides in combination with penicillins (19%). CONCLUSIONS: Antimicrobial prescription increased in Brazilian ICUs during the COVID-19 pandemic, especially during the first months of the epidemic. We identified greater use of broad-spectrum antimicrobials by COVID-19 patients. Overall, the DDD metric overestimated antimicrobial use compared with the DOT metric.

Clinical Characteristics and Outcomes of Patients with Aneurysmal Subarachnoid Hemorrhage: A Prospective Multicenter Study in a Middle-Income Country.

BACKGROUND: Aneurysmal subarachnoid hemorrhage (SAH) is associated with high mortality and long-term functional impairment. Data on clinical management and functional outcomes from developing countries are scarce. We aimed to define patient profiles and clinical practices and evaluate long-term outcomes after SAH in a middle-income country. METHODS: This was a prospective study including consecutive adult patients admitted with SAH to two reference centers in Brazil from January 2016 to February 2020. The primary outcome was functional status at 6 months using the modified Rankin Scale. Mixed multivariable analysis was performed to determine the relationship between clinical variables and functional outcomes. RESULTS: From 471patients analyzed, the median time from symptom onset to arrival at a study center was 4 days (interquartile range 0-9). Median age was 55 years (interquartile range 46-62) and 353 (75%) patients were women. A total of 426 patients (90%) were transferred from nonspecialized general hospitals, initial computed tomography revealed thick hemorrhage in 73% of patients (modified Fisher score of 3 or 4), and 136 (29%) had poor clinical grade (World Federation of Neurological Surgeons score of 4 or 5). A total of 312 (66%) patients underwent surgical clipping, and 119 (25%) underwent endovascular coiling. Only 34 patients (7%) underwent withdrawal or withholding of life-sustaining therapy during their hospital stay, and in-hospital mortality was 24%. A total of 187 (40%) patients had an unfavorable long-term functional outcome (modified Rankin Scale score of 4 to 6). Factors associated with unfavorable outcome were age (adjusted odds ratio [OR] 1.05, 95% confidence interval [CI] 1.03-1.08), hypertension (adjusted OR 1.81, 95% CI 1.04-3.16), poor clinical grade (adjusted OR 4.92, 95% CI 2.85-8.48), external ventricular drain (adjusted OR 3.8, 95% CI 2.31-6.24), postoperative deterioration (adjusted OR 2.33, 95% CI 1.32-4.13), cerebral infarction (adjusted OR 3.16, 95% CI 1.81-5.52), rebleeding (adjusted OR 2.95, 95% CI 1.13-7.69), and sepsis (adjusted OR 2.68, 95% CI 1.42-5.05). CONCLUSIONS: Our study demonstrated that SAH management in a middle-income country diverges significantly from published cohorts and current guidelines, despite comparable clinical profiles on presentation and admission to high-volume referral centers. Earlier aneurysm occlusion and increased use of endovascular therapy could potentially reduce modifiable in-hospital complications and improve functional outcomes in Brazil.

Effects of Tranexamic Acid in Patients with Subarachnoid Hemorrhage in Brazil: A Prospective Observational Study with Propensity Score Analysis.

BACKGROUND: Rebleeding from a ruptured aneurysm increases the risk of unfavorable outcomes after subarachnoid hemorrhage (SAH) and is prevented by early aneurysm occlusion. The role of antifibrinolytics before aneurysm obliteration remains controversial. We investigated the effects of tranexamic acid on long-term functional outcomes of patients with aneurysmal SAH (aSAH). METHODS: This was a single-center, prospective, observational study conducted in a high-volume tertiary hospital in a middle-income country from December 2016 to February 2020. We included all consecutive patients with aSAH who either received or did not receive tranexamic acid (TXA) treatment. Multivariate logistic regression analysis using propensity score was used to evaluate the association of TXA use with long-term functional outcomes, measured by the modified Rankin Scale (mRS) at 6 months. RESULTS: A total of 230 patients with aSAH were analyzed. The median (interquartile range) age was 55 (46-63) years, 72% were women, 75% presented with good clinical grade (World Federation of Neurological Surgeons grade 1-3), and 83% had a Fisher scale of 3 or 4. Around 80% of patients were admitted up to 72 h from ictus. The aneurysm occlusion method was surgical clipping in 80% of the patients. A total of 129 patients (56%) received TXA. In multivariable logistic regression using inverse probability treatment weighting, the long-term rate of unfavorable outcomes (modified Rankin scale 4-6) was the same in the TXA and non-TXA groups (61 [48%] in TXA group vs. 33 [33%] in non-TXA group; odds ratio [OR] 1.39, 95% confidence interval [CI] 0.67-2.92; p = 0.377). The TXA group had higher in-hospital mortality (33 vs. 11% in non-TXA group; OR 4.13, 95% CI 1.55-12.53, p = 0.007). There were no differences between the groups concerning intensive care unit length of stay (16 ± 11.22 days in TXA group vs. 14 ± 9.24 days in non-TXA group; p = 0.2) or hospital (23 ± 13.35 days in TXA group vs. 22 ± 13.36 days in non-TXA group; p = 0.9). There was no difference in the rates of rebleeding (7.8% in TXA group vs. 8.9% in non-TXA group; p = 0.31) or delayed cerebral ischemia (27% in TXA group vs. 19% in non-TXA group; p = 0.14). For the propensity-matched analysis, 128 individuals were selected (64 in TXA group and 64 in non-TXA group), and the rates of unfavorable outcomes at 6 months were also similar between groups (45% in TXA group and 36% in non-TXA group; OR 1.22, 95% CI 0.51-2.89; p = 0.655). CONCLUSIONS: Our findings in a cohort with delayed aneurysm treatment reinforce previous data that TXA use before aneurysm occlusion does not improve functional outcomes in aSAH.

Neurological Pupil Index and Delayed Cerebral Ischemia after Subarachnoid Hemorrhage: A Retrospective Multicentric Study.

BACKGROUND: Delayed cerebral ischemia (DCI) occurs in around 30% of patients suffering from nontraumatic subarachnoid hemorrhage (SAH) and is associated with poor neurological outcome. Whether the Neurological Pupil index (NPi) derived from the automated pupillometry could help to diagnose the occurrence of DCI remains unknown. The aim of this study was to investigate the association of NPi with the occurrence of DCI in patients with SAH. METHODS: This was a multicenter, retrospective cohort study of consecutive patients with SAH admitted to the intensive care units of five hospitals between January 2018 and December 2020 who underwent daily NPi recordings (every 8 h) during the first 10 days of admission. DCI was diagnosed according to standard definitions (in awake patients) or based on neuroimaging and neuromonitoring (in sedated or unconscious patients). An NPi < 3 was defined as abnormal. The primary outcome of the study was to assess the time course of daily NPi between patients with DCI and patients without DCI. Secondary outcome included the number of patients who had an NPi < 3 before DCI. RESULTS: A total of 210 patients were eligible for the final analysis; DCI occurred in 85 (41%) patients. Patients who developed DCI had similar values of mean and worst daily NPi over time when compared with patients without DCI. Patients with DCI had a higher proportion of at least one NPi < 3 at any moment before DCI when compared with others (39/85, 46% vs. 35/125, 38%, p = 0.009). Similarly, the worst NPi before DCI diagnosis was lower in the DCI group when compared with others (3.1 [2.5-3.8] vs. 3.7 [2.7-4.1], p = 0.05). In the multivariable logistic regression analysis, the presence of NPi < 3 was not independently associated with the development of DCI (odds ratio 1.52 [95% confidence interval 0.80-2.88]). CONCLUSIONS: In this study, NPi measured three times a day and derived from the automated pupillometry had a limited value for the diagnosis of DCI in patients with SAH.

Platelet activation and platelet-monocyte aggregate formation trigger tissue factor expression in patients with severe COVID-19.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an emergent pathogen responsible for the coronavirus disease 2019 (COVID-19). Since its emergence, the novel coronavirus has rapidly achieved pandemic proportions causing remarkably increased morbidity and mortality around the world. A hypercoagulability state has been reported as a major pathologic event in COVID-19, and thromboembolic complications listed among life-threatening complications of the disease. Platelets are chief effector cells of hemostasis and pathological thrombosis. However, the participation of platelets in the pathogenesis of COVID-19 remains elusive. This report demonstrates that increased platelet activation and platelet-monocyte aggregate formation are observed in severe COVID-19 patients, but not in patients presenting mild COVID-19 syndrome. In addition, exposure to plasma from severe COVID-19 patients increased the activation of control platelets ex vivo. In our cohort of COVID-19 patients admitted to the intensive care unit, platelet-monocyte interaction was strongly associated with tissue factor (TF) expression by the monocytes. Platelet activation and monocyte TF expression were associated with markers of coagulation exacerbation as fibrinogen and D-dimers, and were increased in patients requiring invasive mechanical ventilation or patients who evolved with in-hospital mortality. Finally, platelets from severe COVID-19 patients were able to induce TF expression ex vivo in monocytes from healthy volunteers, a phenomenon that was inhibited by platelet P-selectin neutralization or integrin αIIb/ß3 blocking with the aggregation inhibitor abciximab. Altogether, these data shed light on new pathological mechanisms involving platelet activation and platelet-dependent monocyte TF expression, which were associated with COVID-19 severity and mortality.

Hospital Length of Stay and 30-Day Mortality Prediction in Stroke: A Machine Learning Analysis of 17,000 ICU Admissions in Brazil.

BACKGROUND: Hospital length of stay and mortality are associated with resource use and clinical severity, respectively, in patients admitted to the intensive care unit (ICU) with acute stroke. We proposed a structured data-driven methodology to develop length of stay and 30-day mortality prediction models in a large multicenter Brazilian ICU cohort. METHODS: We analyzed data from 130 ICUs from 43 Brazilian hospitals. All consecutive adult patients admitted with stroke (ischemic or nontraumatic hemorrhagic) to the ICU from January 2011 to December 2020 were included. Demographic data, comorbidities, acute disease characteristics, organ support, and laboratory data were retrospectively analyzed by a data-driven methodology, which included seven different types of machine learning models applied to training and test sets of data. The best performing models, based on discrimination and calibration measures, are reported as the main results. Outcomes were hospital length of stay and 30-day in-hospital mortality. RESULTS: Of 17,115 ICU admissions for stroke, 16,592 adult patients (13,258 ischemic and 3334 hemorrhagic) were analyzed; 4298 (26%) patients had a prolonged hospital length of stay (> 14 days), and 30-day mortality was 8% (n = 1392). Prolonged hospital length of stay was best predicted by the random forests model (Brier score = 0.17, area under the curve = 0.73, positive predictive value = 0.61, negative predictive value = 0.78). Mortality prediction also yielded the best discrimination and calibration through random forests (Brier score = 0.05, area under the curve = 0.90, positive predictive value = 0.66, negative predictive value = 0.94). Among the 20 strongest contributor variables in both models were (1) premorbid conditions (e.g., functional impairment), (2) multiple organ dysfunction parameters (e.g., hypotension, mechanical ventilation), and (3) acute neurological aspects of stroke (e.g., Glasgow coma scale score on admission, stroke type). CONCLUSIONS: Hospital length of stay and 30-day mortality of patients admitted to the ICU with stroke were accurately predicted through machine learning methods, even in the absence of stroke-specific data, such as the National Institutes of Health Stroke Scale score or neuroimaging findings. The proposed methods using general intensive care databases may be used for resource use allocation planning and performance assessment of ICUs treating stroke. More detailed acute neurological and management data, as well as long-term functional outcomes, may improve the accuracy and applicability of future machine-learning-based prediction algorithms.

New perspectives on brain death.

Brain death, or death by neurological criteria (BD/DNC), has been accepted conceptually, medically and legally for decades. Nevertheless, some areas remain controversial or understudied, pointing to a need for focused research to advance the field. Multiple recent contributions have increased our understanding of BD/DNC, solidified our practice and provided guidance where previously lacking. There have also been important developments on a global scale, including in low-to-middle income countries such as in South America. Although variability in protocols and practice still exists, new efforts are underway to reduce inconsistencies and better train practitioners in accurate and sound BD/DNC determination. Various legal challenges have required formal responses from national societies, and the American Academy of Neurology has filled this void with much needed guidance. Questions remain regarding concepts such as 'whole brain' versus 'brainstem' death, and the intersection of BD/DNC and rubrics of medical futility. These concepts are the subject of this review.

Systemic Severity and Organ Dysfunction in Subarachnoid Hemorrhage: A Large Retrospective Multicenter Cohort Study.

BACKGROUND AND PURPOSE: Acute physiologic derangements and multiple organ dysfunction are common after subarachnoid hemorrhage. We aimed to evaluate the simplified acute physiology score 3 (SAPS-3) and the sequential organ failure assessment (SOFA) scores for the prediction of in-hospital mortality in a large multicenter cohort of SAH patients. METHODS: This was a retrospective analysis of prospectively collected data from 45 ICUs in Brazil, during 2014 and 2015. Patients admitted with non-traumatic subarachnoid hemorrhage (SAH) were included. Clinical and outcome data were retrieved from an electronic ICU quality registry. SAPS-3 and SOFA scores, without the neurological components (i.e., nSAPS-3 and nSOFA, respectively) were recorded, as well as the World Federation of Neurological Surgeons (WFNS) scale. We used multilevel logistic regression analysis to identify factors associated with in-hospital mortality. We evaluated performance using the area under the receiver operating characteristic curve (AUROC), as well as calibration belts and precision-recall plots. RESULTS: The study included 997 patients, from which 426 (43%) had poor clinical grade (WFNS 4 or 5) and in-hospital mortality was 34%. Median nSAPS-3 and nSOFA score at admission were 46 (IQR: 38-55) and 2 (0-5), respectively. Non-survivors were older, had higher nSAPS-3 and nSOFA, and more often poor grade. After adjustment for age, poor grade and withdrawal of life sustaining therapies, multivariable analysis identified nSAPS-3 and nSOFA score as independent clinical predictors of in-hospital mortality. The AUROC curve that included nSAPS-3 and nSOFA scores significantly improved the already good discrimination and calibration of age and WFNS to predict in-hospital mortality (AUROC: 0.89 for the full final model vs. 0.85 for age and WFNS; P < 0.0001). CONCLUSIONS: nSAPS-3 and nSOFA scores were independently associated with in-hospital mortality after SAH. The addition of these scores improved early prediction of hospital mortality in our cohort and should be integrated to other specific prognostic indices in the early assessment of SAH.

The Global Consortium Study of Neurological Dysfunction in COVID-19 (GCS-NeuroCOVID): Development of Case Report Forms for Global Use.

Since its original report in January 2020, the coronavirus disease 2019 (COVID-19) due to Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2) infection has rapidly become one of the deadliest global pandemics. Early reports indicate possible neurological manifestations associated with COVID-19, with symptoms ranging from mild to severe, highly variable prevalence rates, and uncertainty regarding causal or coincidental occurrence of symptoms. As neurological involvement of any systemic disease is frequently associated with adverse effects on morbidity and mortality, obtaining accurate and consistent global data on the extent to which COVID-19 may impact the nervous system is urgently needed. To address this need, investigators from the Neurocritical Care Society launched the Global Consortium Study of Neurological Dysfunction in COVID-19 (GCS-NeuroCOVID). The GCS-NeuroCOVID consortium rapidly implemented a Tier 1, pragmatic study to establish phenotypes and prevalence of neurological manifestations of COVID-19. A key component of this global collaboration is development and application of common data elements (CDEs) and definitions to facilitate rigorous and systematic data collection across resource settings. Integration of these elements is critical to reduce heterogeneity of data and allow for future high-quality meta-analyses. The GCS-NeuroCOVID consortium specifically designed these elements to be feasible for clinician investigators during a global pandemic when healthcare systems are likely overwhelmed and resources for research may be limited. Elements include pediatric components and translated versions to facilitate collaboration and data capture in Latin America, one of the epicenters of this global outbreak. In this manuscript, we share the specific data elements, definitions, and rationale for the adult and pediatric CDEs for Tier 1 of the GCS-NeuroCOVID consortium, as well as the translated versions adapted for use in Latin America. Global efforts are underway to further harmonize CDEs with other large consortia studying neurological and general aspects of COVID-19 infections. Ultimately, the GCS-NeuroCOVID consortium network provides a critical infrastructure to systematically capture data in current and future unanticipated disasters and disease outbreaks.

The Effect of Packed Red Blood Cell Transfusion on Cerebral Oxygenation and Metabolism After Subarachnoid Hemorrhage.

BACKGROUND: Anemia adversely affects cerebral oxygenation and metabolism after subarachnoid hemorrhage (SAH) and is also associated with poor outcome. There is limited evidence to support the use of packed red blood cell (PRBC) transfusion to optimize brain homeostasis after SAH. The aim of this study was to investigate the effect of transfusion on cerebral oxygenation and metabolism in patients with SAH. METHODS: This was a prospective observational study in a neurological intensive care unit of a university hospital. Nineteen transfusions were studied in 15 consecutive patients with SAH that underwent multimodality monitoring (intracranial pressure, brain tissue oxygen, and cerebral microdialysis). Data were collected at baseline and for 12 h after transfusion. The relationship between hemoglobin (Hb) change and lactate/pyruvate ratio (LPR) orbrain tissue oxygen (PbtO2) was tested using univariate and multivariable analyses. RESULTS: PRBC transfusion was administered on the median post-bleed day 8. The average Hb concentration at baseline was 8.1 g/dL and increased by 2.2 g/dL after transfusion. PbtO2 increased between hours 2 and 4 post-transfusion and this increase was maintained until hour 10. LPR did not change significantly during the 12-h monitoring period. After adjusting for SpO2, cerebral perfusion pressure, and LPR, the change in Hb concentration was independently and positively associated with a change in PbtO2 (adjusted b estimate = 1.39 [95% confidence interval 0.09-2.69]; P = 0.04). No relationship between the change in Hb concentration and LPR was found. CONCLUSIONS: PRBC transfusion resulted in PbtO2 improvement without a clear effect on cerebral metabolism prior to SAH.

Systemic glucose variability predicts cerebral metabolic distress and mortality after subarachnoid hemorrhage: a retrospective observational study.

INTRODUCTION: Cerebral glucose metabolism and energy production are affected by serum glucose levels. Systemic glucose variability has been shown to be associated with poor outcome in critically ill patients. The objective of this study was to assess whether glucose variability is associated with cerebral metabolic distress and outcome after subarachnoid hemorrhage. METHODS: A total of 28 consecutive comatose patients with subarachnoid hemorrhage, who underwent cerebral microdialysis and intracranial pressure monitoring, were studied. Metabolic distress was defined as lactate/pyruvate ratio (LPR) >40. The relationship between daily glucose variability, the development of cerebral metabolic distress and hospital outcome was analyzed using a multivariable general linear model with a logistic link function for dichotomized outcomes. RESULTS: Daily serum glucose variability was expressed as the standard deviation (SD) of all serum glucose measurements. General linear models were used to relate this predictor variable to cerebral metabolic distress and mortality at hospital discharge. A total of 3,139 neuromonitoring hours and 181 days were analyzed. After adjustment for Glasgow Coma Scale (GCS) scores and brain glucose, SD was independently associated with higher risk of cerebral metabolic distress (adjusted odds ratio = 1.5 (1.1 to 2.1), P = 0.02). Increased variability was also independently associated with in hospital mortality after adjusting for age, Hunt Hess, daily GCS and symptomatic vasospasm (P = 0.03). CONCLUSIONS: Increased systemic glucose variability is associated with cerebral metabolic distress and increased hospital mortality. Therapeutic approaches that reduce glucose variability may impact on brain metabolism and outcome after subarachnoid hemorrhage.

Fluid responsiveness and brain tissue oxygen augmentation after subarachnoid hemorrhage.

BACKGROUND: The objective of this study was to investigate the relationship between cardiac index (CI) response to a fluid challenge and changes in brain tissue oxygen pressure (PbtO(2)) in patients with subarachnoid hemorrhage (SAH). METHODS: Prospective observational study was conducted in a neurological intensive care unit of a university hospital. Fifty-seven fluid challenges were administered to ten consecutive comatose SAH patients that underwent multimodality monitoring of CI, intracranial pressure (ICP), and PbtO(2), according to a standardized fluid management protocol. RESULTS: The relationship between CI and PbtO(2) was analyzed with logistic regression utilizing generalized estimating equations. Of the 57 fluid boluses analyzed, 27 (47 %) resulted in a ≥ 10 % increase in CI. Median absolute (+5.8 vs. +1.3 mmHg) and percent (20.7 vs. 3.5 %) changes in PbtO(2) were greater in CI responders than in non-responders within 30 min after the end of the fluid bolus infusion. In a multivariable model, a CI response was independently associated with PbtO(2) response (adjusted odds ratio 21.5, 95 % CI 1.4-324, P = 0.03) after adjusting for mean arterial pressure change and end-tidal CO(2). Stroke volume variation showed a good ability to predict CI and PbtO(2) response with areas under the ROC curve of 0.86 and 0.81 with the best cut-off values of 9 % for both responses. CONCLUSION: Bolus fluid resuscitation resulting in augmentation of CI can improve cerebral oxygenation after SAH.

The Association Between Prepandemic ICU Performance and Mortality Variation in COVID-19: A Multicenter Cohort Study of 35,619 Critically Ill Patients.

BACKGROUND: During the COVID-19 pandemic, ICUs remained under stress and observed elevated mortality rates and high variations of outcomes. A knowledge gap exists regarding whether an ICU performing best during nonpandemic times would still perform better when under high pressure compared with the least performing ICUs. RESEARCH QUESTION: Does prepandemic ICU performance explain the risk-adjusted mortality variability for critically ill patients with COVID-19? STUDY DESIGN AND METHODS: This study examined a cohort of adults with real-time polymerase chain reaction-confirmed COVID-19 admitted to 156 ICUs in 35 hospitals from February 16, 2020, through December 31, 2021, in Brazil. We evaluated crude and adjusted in-hospital mortality variability of patients with COVID-19 in the ICU during the pandemic. Association of baseline (prepandemic) ICU performance and in-hospital mortality was examined using a variable life-adjusted display (VLAD) during the pandemic and a multivariable mixed regression model adjusted by clinical characteristics, interaction of performance with the year of admission, and mechanical ventilation at admission. RESULTS: Thirty-five thousand six hundred nineteen patients with confirmed COVID-19 were evaluated. The median age was 52 years, median Simplified Acute Physiology Score 3 was 42, and 18% underwent invasive mechanical ventilation. In-hospital mortality was 13% and 54% for those receiving invasive mechanical ventilation. Adjusted in-hospital mortality ranged from 3.6% to 63.2%. VLAD in the most efficient ICUs was higher than the overall median in 18% of weeks, whereas VLAD was 62% and 84% in the underachieving and least efficient groups, respectively. The least efficient baseline ICU performance group was associated independently with increased mortality (OR, 2.30; 95% CI, 1.45-3.62) after adjusting for patient characteristics, disease severity, and pandemic surge. INTERPRETATION: ICUs caring for patients with COVID-19 presented substantial variation in risk-adjusted mortality. ICUs with better baseline (prepandemic) performance showed reduced mortality and less variability. Our findings suggest that achieving ICU efficiency by targeting improvement in organizational aspects of ICUs may impact outcomes, and therefore should be a part of the preparedness for future pandemics.

Development and validation of a machine learning model to predict the use of renal replacement therapy in 14,374 patients with COVID-19.

PURPOSE: To develop a model to predict the use of renal replacement therapy (RRT) in COVID-19 patients. MATERIALS AND METHODS: Retrospective analysis of multicenter cohort of intensive care unit (ICU) admissions of Brazil involving COVID-19 critically adult patients, requiring ventilatory support, admitted to 126 Brazilian ICUs, from February 2020 to December 2021 (development) and January to May 2022 (validation). No interventions were performed. RESULTS: Eight machine learning models' classifications were evaluated. Models were developed using an 80/20 testing/train split ratio and cross-validation. Thirteen candidate predictors were selected using the Recursive Feature Elimination (RFE) algorithm. Discrimination and calibration were assessed. Temporal validation was performed using data from 2022. Of 14,374 COVID-19 patients with initial respiratory support, 1924 (13%) required RRT. RRT patients were older (65 [53-75] vs. 55 [42-68]), had more comorbidities (Charlson's Comorbidity Index 1.0 [0.00-2.00] vs 0.0 [0.00-1.00]), had higher severity (SAPS-3 median: 61 [51-74] vs 48 [41-58]), and had higher in-hospital mortality (71% vs 22%) compared to non-RRT. Risk factors for RRT, such as Creatinine, Glasgow Coma Scale, Urea, Invasive Mechanical Ventilation, Age, Chronic Kidney Disease, Platelets count, Vasopressors, Noninvasive Ventilation, Hypertension, Diabetes, modified frailty index (mFI) and Gender, were identified. The best discrimination and calibration were found in the Random Forest (AUC [95%CI]: 0.78 [0.75-0.81] and Brier's Score: 0.09 [95%CI: 0.08-0.10]). The final model (Random Forest) showed comparable performance in the temporal validation (AUC [95%CI]: 0.79 [0.75-0.84] and Brier's Score, 0.08 [95%CI: 0.08-0.1]). CONCLUSIONS: An early ML model using easily available clinical and laboratory data accurately predicted the use of RRT in critically ill patients with COVID-19. Our study demonstrates that using ML techniques is feasible to provide early prediction of use of RRT in COVID-19 patients.

Early neurological deterioration after subarachnoid haemorrhage: risk factors and impact on outcome.

BACKGROUND: Early neurological deterioration occurs frequently after subarachnoid haemorrhage (SAH). The impact on hospital course and outcome remains poorly defined. METHODS: We identified risk factors for worsening on the Hunt-Hess grading scale within the first 24 h after admission in 609 consecutively admitted aneurysmal SAH patients. Admission risk factors and the impact of early worsening on outcome was evaluated using multivariable analysis adjusting for age, gender, admission clinical grade, admission year and procedure type. Outcome was evaluated at 12 months using the modified Rankin Scale (mRS). RESULTS: 211 patients worsened within the first 24 h of admission (35%). In a multivariate adjusted model, early worsening was associated with older age (OR 1.02, 95% CI 1.001 to 1.03; p=0.04), the presence of intracerebral haematoma on initial CT scan (OR 2.0, 95% CI 1.2 to 3.5; p=0.01) and higher SAH and intraventricular haemorrhage sum scores (OR 1.05, 95% CI 1.03 to 1.08 and 1.1, 95% CI 1.01 to 1.2; p<0.001 and 0.03, respectively). Early worsening was associated with more hospital complications and prolonged length of hospital stay and was an independent predictor of death (OR 12.1, 95% CI 5.7 to 26.1; p<0.001) and death or moderate to severe disability (mRS 4-6, OR 8.4, 95% CI 4.9 to 14.5; p=0.01) at 1 year. CONCLUSIONS: Early worsening after SAH occurs in 35% of patients, is predicted by clot burden and is associated with mortality and poor functional outcome at 1 year.

Reduced brain/serum glucose ratios predict cerebral metabolic distress and mortality after severe brain injury.

BACKGROUND: The brain is dependent on glucose to meet its energy demands. We sought to evaluate the potential importance of impaired glucose transport by assessing the relationship between brain/serum glucose ratios, cerebral metabolic distress, and mortality after severe brain injury. METHODS: We studied 46 consecutive comatose patients with subarachnoid or intracerebral hemorrhage, traumatic brain injury, or cardiac arrest who underwent cerebral microdialysis and intracranial pressure monitoring. Continuous insulin infusion was used to maintain target serum glucose levels of 80-120 mg/dL (4.4-6.7 mmol/L). General linear models of logistic function utilizing generalized estimating equations were used to relate predictors of cerebral metabolic distress (defined as a lactate/pyruvate ratio [LPR] ≥ 40) and mortality. RESULTS: A total of 5,187 neuromonitoring hours over 300 days were analyzed. Mean serum glucose was 133 mg/dL (7.4 mmol/L). The median brain/serum glucose ratio, calculated hourly, was substantially lower (0.12) than the expected normal ratio of 0.40 (brain 2.0 and serum 5.0 mmol/L). In addition to low cerebral perfusion pressure (P = 0.05) and baseline Glasgow Coma Scale score (P < 0.0001), brain/serum glucose ratios below the median of 0.12 were independently associated with an increased risk of metabolic distress (adjusted OR = 1.4 [1.2-1.7], P < 0.001). Low brain/serum glucose ratios were also independently associated with in-hospital mortality (adjusted OR = 6.7 [1.2-38.9], P < 0.03) in addition to Glasgow Coma Scale scores (P = 0.029). CONCLUSIONS: Reduced brain/serum glucose ratios, consistent with impaired glucose transport across the blood brain barrier, are associated with cerebral metabolic distress and increased mortality after severe brain injury.