Journal of Intensive Care Medicine最新文献

IntroductionGlucocorticoids are commonly used in pediatric critical illness and may lead to subsequent adrenal insufficiency, causing morbidity among pediatric intensive care unit (PICU) survivors. We aimed to determine the prevalence of and risk factors for adrenal insufficiency among children who received glucocorticoids during PICU admission.MethodsWe conducted a retrospective cohort study using structured medical record review to determine the prevalence of adrenal insufficiency and clinical characteristics of PICU patients 0-18 years old who received enteral and/or parenteral glucocorticoids. Patients were consecutively admitted to an academic tertiary referral PICU over 2 years.ResultsAmong 530 patients who received glucocorticoids, 12 (2.3%) were diagnosed with adrenal insufficiency at a median of 55 (IQR 8-156) days after initial glucocorticoid exposure. Unadjusted analyses showed that patients with adrenal insufficiency were younger (median 0.5 vs 2 years, p = .020), had a longer PICU stay (79 vs 4 days, p < .001) and hospital stay (96 vs 6 days, p < .001), and had a lower survival rate at 1 year after PICU discharge (75% vs 94%, p = .033). There were no significant differences in sex, race/ethnicity, illness severity, or diagnostic categories. Patients with adrenal insufficiency were more likely to have received glucocorticoids for hyperinflammation (21% vs 8%) and less likely for reactive airway disease (10% vs 26%) (p = .036), had a higher median total hydrocortisone equivalent dose (2508 vs 480 mg, p = .007), and were more likely to have had a steroid taper (48% vs 24%, p = .003). Multivariable logistic regression showed no significant associations between clinical characteristics and the diagnosis of adrenal insufficiency.ConclusionsAmong PICU patients who received glucocorticoids, 2.3% were subsequently diagnosed with adrenal insufficiency. We identified potential risk factors for adrenal insufficiency after glucocorticoid use in the PICU, which warrant future study to better delineate and mitigate adrenal insufficiency's contribution to morbidity and mortality among critically ill children.

PurposeThe aim of this study was to describe seasonal variation in the incidence and outcomes of sepsis in the United States.MethodsThis is a retrospective study using National Inpatient Sample database from 2017-2019. Adult sepsis patients were identified based on the CMS SEP-1 measure codes. Monthly sepsis incidence, in-hospital mortality, and organ failure patterns were analyzed. Multivariable logistic regression was used to assess in-hospital mortality by month. We also analyzed the monthly variation in each type of organ failure to uncover patterns that could account for the observed differences in sepsis incidence and outcomes.Main ResultsThere were 57,019,369 hospitalizations due to sepsis during the study period. The incidence of sepsis hospitalizations was highest in January. January also had the highest in-hospital mortality rate (10.66%), while July had the lowest (8.66%). A multivariable logistic regression analysis showed that January had a significantly higher mortality rate compared to July (odds ratio of 1.22, p < 0.001). The relationship between month and in-hospital mortality for sepsis followed a U-shaped pattern (from January to December), both in raw and adjusted analysis. Respiratory failure similarly followed the U-shaped pattern, with January having the highest incidence, and July and August the lowest. Other organ failures showed consistent patterns throughout the year. The relationship between sepsis due to pneumonia was also U-shaped, especially in the Southern region.ConclusionsThis study revealed a U-shaped relationship between both incidence and in-hospital mortality of sepsis, and month throughout the year, with a peak during winter months. Respiratory failure significantly increased in winter, while other organ failures remained constant throughout the year. These data suggest that respiratory infection and respiratory failure appear to mediate the seasonal variation observed in sepsis incidence and mortality, respectively.

BackgroundExtracorporeal Membrane Oxygenation (ECMO) is a critical support system for patients with acute and severe cardiac and respiratory failure. This study investigates the impact of different patient body weight categories on the mortality rates of patients undergoing ECMO support.MethodsUsing the Nationwide Sample (NIS) database and ICD-10 codes for 2016 to 2020 in adults over age 18, we evaluated total mortality based on weight categories compared to normal weights using univariate and multivariate analyses.ResultsA total population of 47 990 patients underwent ECMO insertion with a mean age of 52.6 years. Total mortality was 45.7%. Patients with cachexia, overweight, and obesity had similar mortality to normal-weight patients. (Cachexia: 43.75%, normal weight: 46.30%, p = .60, OR = 0.90, 95% CI: 0.61-1.33, overweight 42.31%, p = .69, OR = 0.85, 95% CI: 0.38-1.89, and obesity 45.73%, p = .73, OR = 0.98, 95% CI: 0.85-1.12). However, morbid obesity had the lowest mortality in the univariate analysis (41.89%, p = .01, OR = 0.84, 95% CI: 0.73-0.96) but was not significant in the multivariate analysis (p = .66, OR: 0.97, CI: 0.83-1.12). Separating peripheral veno-arterial versus veno-venous ECMO showed similar results with similar mortalities based on weight categories.ConclusionsOur data suggest that the 'obesity paradox' does not exist in ECMO-treated patients, with no effect of weight on total mortality . Further research is necessary to understand the underlying factors contributing to these outcomes.

Post-intensive care syndrome (PICS) is a clinical syndrome characterized by new or worsening changes in mental health, cognition, or physical function that persist following critical illness. The psychiatric domain of PICS encompasses new or worsened psychiatric burdens following critical illness, including post-traumatic stress disorder (PTSD), depression, and anxiety. Many of the established predisposing and precipitating factors for the psychiatric domain of PICS are commonly found in the setting of critical illness, including mechanical ventilation (MV), exposure to sedating medications, and physical restraint. Importantly, previous psychiatric history is a strong risk factor for the development of the psychiatric domain of PICS and should be considered when screening patients to diagnose psychiatric impairment and interventions. Delirium has been associated with psychiatric symptoms following ICU admission, therefore prevention warrants careful consideration. Dexmedetomidine has been shown to have the lowest risk for development of delirium when compared to other sedatives and has been the only sedative studied in relation to the psychiatric domain of PICS. Nocturnal dexmedetomidine and intensive care unit (ICU) diaries have been associated with decreased psychiatric burden after ICU discharge. Studies evaluating the impact of other intra-ICU practices on the development of the psychiatric domain of PICS, including the ABCDEF bundle, depth of sedation, and daily spontaneous awakening trials, have been limited and inconclusive. The psychiatric domain of PICS is difficult to treat and may be less responsive to multidisciplinary post-discharge programs and targeted interventions than the cognitive and physical domains of PICS. Given the high morbidity associated with the psychiatric domain of PICS, intensivists should familiarize themselves with the risk factors and intra-ICU interventions that can mitigate this important and under-recognized condition.

Background and ObjectiveHealthcare professionals may be able to anticipate more accurately a patient's timing of death and assess their possibility of recovery by implementing a real-time clinical decision support system. Using such a tool, the healthcare system can better understand a patient's condition and make more informed judgements about distributing limited resources. This scoping review aimed to analyze various death prediction AI (Artificial Intelligence) algorithms that have been used in ICU (Intensive Care Unit) patient populations.MethodsThe search strategy of this study involved keyword combinations of outcome and patient setting such as mortality, survival, ICU, terminal care. These terms were used to perform database searches in MEDLINE, Embase, and PubMed up to July 2022. The variables, characteristics, and performance of the identified predictive models were summarized. The accuracy of the models was compared using their Area Under the Curve (AUC) values.ResultsDatabases search yielded an initial pool of 8271 articles. A two-step screening process was then applied: first, titles and abstracts were reviewed for relevance, reducing the pool to 429 articles. Next, a full-text review was conducted, further narrowing down the selection to 400 key studies. Out of 400 studies on different tools or models for prediction of mortality in ICUs, 16 papers focused on AI-based models which were ultimately included in this study that have deployed different AI-based and machine learning models to make a prediction about negative patient outcome. The accuracy and performance of the different models varied depending on the patient populations and medical conditions. It was found that AI models compared with traditional tools like SAP3 or APACHE IV score were more accurate in death prediction, with some models achieving an AUC of up to 92.9%. The overall mortality rate ranged from 5% to more than 60% in different studies.ConclusionWe found that AI-based models exhibit varying performance across different patient populations. To enhance the accuracy of mortality prediction, we recommend customizing models for specific patient groups and medical contexts. By doing so, healthcare professionals may more effectively assess mortality risk and tailor treatments accordingly. Additionally, incorporating additional variables-such as genetic information-into new models can further improve their accuracy.

IntroductionIn 2024, a Society of Critical Care Medicine task force updated the pediatric sepsis definition from the presence of suspected or confirmed infection, and a systemic inflammatory response (SIRS) with organ dysfunction, to a novel definition. Our objective is to identify how many patients previously identified as having severe sepsis would continue to meet the new definition.Materials and methodsWe performed a secondary analysis of the Phenotyping Sepsis-Induced Multiple Organ Failure cohort of 401 children with suspected or confirmed infection, two of four SIRS criteria and organ dysfunction enrolled between 2015-2017. We calculated a modified Phoenix Sepsis Criteria Score (mPSC) for participants and compared those with mPSC of greater than or equal to 2 or less than 2 according to the 2024 definition.ResultsOf 401 children, 132 (33%) did not meet mPSC definitions. While children meeting mPSC had more organ dysfunction, the total mortality did not differ. One in 4 children requiring extracorporeal membrane oxygenation and 1 in 4 mortalities did not meet the mPSC definition. In logistic regression models, in the complete cohort, hematologic (OR 4.4, 95% CI: 1.8-10.2, P-value = .001), central nervous system (OR 2.3, 95% CI: 1.0-5.1, P-value = .046) and renal failure (OR: 3.2, 95% CI:1.2-7.9, P-value = .017) predicted mortality; in the mPSC subgroup pulmonary (OR: 3.6, 95% CI:1.3-13.3, P-value = .030) and hematologic failure (OR 5.6, 95% CI: 2.2-14.5, P-value = .0003) were significant predictors. In the mPSC excluded subgroup, only renal failure predicted mortality (OR 9.6, 95% CI 1.1-73.0, P-value = .028).ConclusionsFurther study of the impact of the 2024 data-driven organ dysfunction definition on pediatric sepsis research, patient safety, and clinical benchmarking efforts is warranted.

BackgroundThe incidence of acute kidney injury (AKI) is increased in patients with community-acquired pneumonia (CAP), contributing to poor outcomes in ICUs. Early identification of patients at high risk for AKI is essential for timely intervention. This study aimed to develop a machine learning model for predicting AKI in CAP patients.MethodsPatients with CAP were identified from the MIMIC-IV database using ICD codes. AKI was defined according to the KDIGO criteria. Baseline characteristics, vital signs, laboratory data, comorbidities, and clinical scores were extracted. LASSO regression was applied for feature selection, and eight machine learning models, including logistic regression, k-nearest neighbors, decision tree, random forest, support vector machine, neural network, XGBoost, and LightGBM, were developed. Model performance was evaluated using AUC, sensitivity, specificity, accuracy, recall, F1 score, calibration curves, and decision curve analysis (DCA). SHapley Additive exPlanations (SHAP) were used to interpret the final model. A web-based risk calculator was created for clinical application.ResultsA total of 3213 CAP patients were included, with 2723 (84.8%) developing AKI. XGBoost demonstrated the best performance with an AUC of 0.937 (95% CI: 0.922-0.952), sensitivity of 0.875, specificity of 0.855, accuracy of 0.865 (95% CI: 0.841-0.887), recall of 0.875, and F1 score of 0.866. DCA showed the highest net benefit for XGBoost across various risk thresholds. After recursive feature elimination, a simplified model with seven key variables, including urine output, weight, ventilation, first-day minimum PTT, first-day maximum sodium, first-day minimum heart rate, and first-day maximum temperature, maintained high predictive performance (AUC = 0.925, 95% CI: 0.908-0.941).ConclusionsThe XGBoost model accurately predicted AKI risk in CAP patients, demonstrating robust performance and clinical utility. The web-based calculator offers an accessible tool for individualized risk assessment, supporting early detection and management of AKI in ICUs.

PurposeRight ventricular impairment (RVI) can be alleviated by the initiation of veno-venous extracorporeal membrane oxygenation (V-V ECMO), which enhances gas exchange and allows for less invasive mechanical ventilation. However, the progression of RVI during V-V ECMO remains unclear. This study assesses echocardiographic changes in RVI over a five-day period in twenty acute respiratory distress syndrome (ARDS) patients with V-V ECMO support.Material and MethodsOver a five-day period of V-V ECMO support, we examined echocardiographic markers of RVI, including right and left ventricular end-diastolic area ratio (RVEDA/LVEDA), tricuspid annular plane systolic excursion (TAPSE), tricuspid valve lateral anulus peak systolic velocity (S'), right ventricular fractional area change (FAC), and right ventricular myocardial performance index. Secondary objectives included changes in mechanical power transmitted to the respiratory system, hemodynamics and gas-exchange.ResultsRVEDA/LVEDA ratio remained elevated (0.8 [0.7-0.8] vs 0.7 [0.7-0.9]; p = .986), TAPSE decreased (2.0[1.6-2.5] cm vs 1.7 [1.4-2.2] cm; p = .024) while no changes were observed in S' (16 [13-21] cm/s vs 15 [12-18] cm/s; p = .136) and FAC (38 [27-47] % vs 36 [29-43] %; p = .627). The right ventricular myocardial performance index improved (0.74 [0.45-1.00] vs 0.51 [0.42-0.80]; p = .004). Lung mechanical power was significantly reduced due to a decrease in lung elastic and resistive components.ConclusionsDespite preserved longitudinal function and improved global performance, RVI persisted in severe ARDS patients on V-V ECMO, as indicated by the RVEDA/LVEDA ratio. These findings suggest that mechanisms beyond hypoxemia, hypercapnia and the invasiveness of mechanical ventilation contribute to RVI in these patients.Trial registrationThis trial was registered with the German Clinical Trials Register (DRKS00028584) on March 28, 2022. https://drks.de/search/en/trial/DRKS00028584.