A new severity of illness scale using a subset of Acute Physiology And Chronic Health Evaluation data elements shows comparable predictive accuracy

Impact of the triglyceride-glucose index on 28-day mortality in non-diabetic critically Ill patients with sepsis: a retrospective cohort analysis

INTRODUCTION

Sepsis is a life-threatening condition that poses a globally high mortality rate. Identifying risk factors is crucial. Insulin resistance and the TYG index, associated with metabolic disorders, may play a role. This study explores their correlation with mortality in non-diabetic septic patients.

METHODS

This retrospective cohort study used data from the MIMIC-IV (version 2.1) database, which includes over 50,000 ICU admissions from 2008 to 2019 at Beth Israel Deaconess Medical Center in Boston. We included adult patients with sepsis who were admitted to the intensive care unit in the study. The primary outcome was to evaluate the ability of TYG to predict death at 28-day of hospital admission in patients with sepsis.

RESULTS

The study included 2213 patients with sepsis, among whom 549 (24.8%) died within 28 days of hospital admission. We observed a non-linear association between TYG and the risk of mortality. Compared to the reference group (lower TYG subgroup), the 28-day mortality increased in the higher TYG subgroup, with a fully adjusted hazard ratio of 2.68 (95% CI: 2.14 to 3.36). The area under the curve (AUC) for TYG was 67.7%, higher than for triglycerides alone (AUC = 64.1%), blood glucose (AUC = 62.4%), and GCS (AUC = 63.6%), and comparable to SOFA (AUC = 69.3%). The final subgroup analysis showed no significant interaction between TYG and each subgroup except for the COPD subgroup (interaction P-values: 0.076-0.548).

CONCLUSION

In our study, TYG can be used as an independent predictor for all-cause mortality due to sepsis within 28 days of hospitalization.

Association of frailty and sarcopenia with short-term mortality in older critically ill patients

BACKGROUND

There is still no study on the use of the SARC-CalF questionnaire for older critically ill patients. Moreover, there is limited evidence on whether a combination of sarcopenia and frailty can provide incremental improvements in risk stratification for older critically ill patients.

METHODS

A total of 653 patients older than 60 years were recruited. We used the clinical frailty scale (CFS) and SARC-CalF questionnaire to assess the frailty status and sarcopenia risk, respectively, of older patients shortly after admission to the ICU. The effect of frailty and sarcopenia risk on ICU mortality and 30-day mortality was evaluated.

RESULTS

A total of 147 (22.5%) patients died in the ICU, and 187 (28.6%) patients died within 30 days after ICU admission. The CFS score was associated with increased ICU mortality [per 1-score increase: odds ratio (OR) = 1.222, 95% confidential interval (CI): 1.003-1.489] and 30-day mortality (per 1-score increase: OR = 1.307, 95% CI: 1.079-1.583). The SARC-CalF score was also associated with increased ICU mortality (per 1-score increase: OR = 1.204, 95% CI: 1.120-1.294) and 30-day mortality (per 1-score increase: OR = 1.247, 95% CI: 1.163-1.337). The addition of the CFS + SARC-CalF score to Acute Physiology and Chronic Health Evaluation (APACHE) II improved discrimination and reclassified ICU and 30-day mortality risk.

CONCLUSIONS

Sarcopenia risk assessed by the SARC-CalF questionnaire provided independent prognostic information for older critically ill patients. A combination of sarcopenia and frailty improved the prediction of mortality for older critically ill patients and thus might be useful in the clinical decision-making process.

Development of a Nomogram Model Based on Lactate-To-Albumin Ratio for Prognostic Prediction in Hospitalized Patients with Intracerebral Hemorrhage

OBJECTIVE

This study aims to develop a nomogram model incorporating lactate-to-albumin ratio (LAR) to predict the prognosis of hospitalized patients with intracerebral hemorrhage (ICH) and demonstrate its excellent predictive performance.

METHODS

A total of 226 patients with ICH from the Medical information mart for intensive care III (MIMIC Ⅲ) database were randomly split into 8:2 ratio training and experimental groups, and 38 patients from the eICU-CRD for external validation. Univariate and multivariate Cox proportional hazards regression analysis was performed to identify independent factors associated with ICH, and multivariate Cox regression was used to construct nomograms for 7-day and 14-day overall survival (OS). The performance of nomogram was verified by the calibration curves, decision curves, and receiver operating characteristic (ROC) curves.

RESULTS

Our study identified LAR, glucose, mean blood pressure, sodium, and ethnicity as independent factors influencing in-hospital prognosis. The predictive performance of our nomogram model for predicting 7-day and 14 -day OS (AUCs: 0.845 and 0.830 respectively) are both superior to Oxford Acute Severity of Illness Score, Simplified acute physiology score II, and SIRS (AUCs: 0.617, 0.620 and 0.591 and AUCs: 0.709, 0.715 and 0.640, respectively) in internal validation, and also demonstrate favorable predictive performance in external validation (AUCs: 0.778 and 0.778 respectively).

CONCLUSIONS

LAR as a novel biomarker is closely associated with an increased risk of in-hospital mortality of patients with ICH. The nomogram model incorporating LAR along with glucose, mean blood pressure, sodium, and ethnicity demonstrate excellent predictive performance for predicting the prognosis of 7- and 14-day OS of hospitalized patients with ICH.

Predictive Value of Multiple Scoring Systems in the Prognosis of Septic Patients with Different Infection Sites: Analysis of the Medical Information Mart for the Intensive Care IV Database

The heterogeneity nature of sepsis is significantly impacted by the site of infection. This study aims to explore the predictive value of multiple scoring systems in assessing the prognosis of septic patients across different infection sites. Data for this retrospective cohort study were extracted from the Medical Information Mart for Intensive Care IV database (MIMIC-IV) (v2.2). Adult patients meeting the criteria for sepsis 3.0 and admitted to the intensive care unit (ICU) were enrolled. Infection sites included were pneumonia, urinary tract infection (UTI), cellulitis, abdominal infection, and bacteremia. The primary outcome assessed was 28-day mortality. The sequential Organ Failure Assessment (SOFA) score, Oxford Acute Severity of Illness Score (OASIS), and Logistic Organ Dysfunction System (LODS) score were compared. Binomial logistic regression analysis was conducted to evaluate the association between these variables and mortality. Additionally, differences in the area under the curve (AUC) of receiver operating characteristic (ROC) among the scoring systems were analyzed. A total of 4721 patients were included in the analysis. The average 28-day mortality rate was 9.4%. Significant differences were observed in LODS, OASIS, and SOFA scores between the 28-day survival and non-survival groups across different infection sites (p < 0.01). In the pneumonia group and abdominal infection group, both the LODS and OASIS scoring systems emerged as independent risk factors for mortality in septic patients (odds ratio [OR]: 1.165, 95% confidence interval [CI]: 1.109-1.224, p < 0.001; OR: 1.047, 95% CI: 1.028-1.065, p < 0.001) (OR: 1.200, 95% CI: 1.091-1.319, p < 0.001; OR: 1.060, 95% CI: 1.025-1.095, p < 0.001). For patients with UTI, the LODS, OASIS, and SOFA scoring systems were identified as independent risk factors for mortality (OR: 1.142, 95% CI: 1.068-1.220, p < 0.001; OR: 1.062, 95% CI: 1.037-1.087, p < 0.001; OR: 1.146, 95% CI: 1.046-1.255, p = 0.004), with the AUC of LODS score and OASIS significantly higher than that of the SOFA score (p = 0.006). Among patients with cellulitis, the OASIS and SOFA scoring systems were identified as independent risk factors for mortality (OR: 1.055, 95% CI: 1.007-1.106, p = 0.025; OR: 1.187, 95% CI: 1.005-1.403, p = 0.044), with no significant difference in prognosis prediction observed (p = 0.243). In the bacteremia group, the LODS scoring system was identified as an independent risk factor for mortality (OR: 1.165, 95% CI: 1.109-1.224, p < 0.001). The findings suggest that LODS scores offer better prognostic accuracy for predicting the mortality risk in septic patients with pneumonia, abdominal infections, bacteremia, and UTI compared to SOFA scores.

Daily variation in blood glucose levels during continuous enteral nutrition in patients on the intensive care unit: a retrospective observational study

BACKGROUND

The circadian timing system coordinates daily cycles in physiological functions, including glucose metabolism and insulin sensitivity. Here, the aim was to characterise the 24-h variation in glucose levels in critically ill patients during continuous enteral nutrition after controlling for potential sources of bias.

METHODS

Time-stamped clinical data from adult patients who stayed in the Intensive Care Unit (ICU) for at least 4 days and received enteral nutrition were extracted from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Linear mixed-effects and XGBoost modelling were used to determine the effect of time of day on blood glucose values.

FINDINGS

In total, 207,647 glucose measurements collected during enteral nutrition were available from 6,929 ICU patients (3,948 males and 2,981 females). Using linear mixed-effects modelling, time of day had a significant effect on blood glucose levels (p < 0.001), with a peak of 9.6 [9.5-9.6; estimated marginal means, 95% CI] mmol/L at 10:00 in the morning and a trough of 8.6 [8.5-8.6] mmol/L at 02:00 at night. A similar impact of time of day on glucose levels was found with the XGBoost regression model.

INTERPRETATION

These results revealed marked 24-h variation in glucose levels in ICU patients even during continuous enteral nutrition. This 24-h pattern persists after adjustment for potential sources of bias, suggesting it may be the result of endogenous biological rhythmicity.

FUNDING

This work was supported by a VENI grant from the Netherlands Organisation for Health Research and Development (ZonMw), an institutional project grant, and by the Dutch Research Council (NWO).

Prognostic value of the systemic immune-inflammation index in critically ill elderly patients with hip fracture: evidence from MIMIC (2008-2019)

Background

The systemic immune-inflammation index (SII) showed an extensive link between immunological dysfunction and the activation of systemic inflammation. Several studies have confirmed the application of SII to orthopedic diseases. However, the significance of SII in critically ill elderly individuals with hip fracture who require intensive care unit (ICU) admission is not yet known. This study centered on exploring the relationship between SII and clinical outcomes among critically ill elderly hip fracture individuals.

Methods

The study centered around elderly patients experiencing severe illness following hip fractures and requiring admission to the ICU. These patients from the MIMIC-IV database formed the basis of this study's cohort. We stratified them into quartiles according to their SII levels. The results involved the mortality at 30 days and 1 year post-admission. Then we employ Cox proportional hazards regression analysis as well as restricted cubic splines to explore the association between the SII and clinical results in critically ill elderly patients with hip fracture.

Results

The study encompassed 991 participants, among whom 63.98% identified as females. Notably, the mortality rates attributed to any cause within 30 days and 1 year after hospitalization stood at 19.68 and 33.40%, respectively. The multivariate Cox proportional hazards model disclosed a significant correlation between an elevated SII and all-cause mortality. Following adjustments for confounding variables, individuals with a high SII showed a notable correlation with 30-day mortality [adjusted hazard ratio (HR), 1.065; 95% confidence interval (CI), 1.044-1.087; p < 0.001] and 1-year mortality (adjusted HR, 1.051; 95% CI, 1.029-1.074; p < 0.001). Furthermore, the analysis of restricted cubic splines demonstrated a progressive increase in the risk of all-cause death as the SII value rose.

Conclusion

Among critically ill elderly patients with hip fracture, the SII exhibits a non-linear association that positively correlates with both 30-day and 1-year all-cause mortality rates. The revelation indicates that the SII may play a vital role in identifying patients with hip fractures who face an escalated risk of mortality due to any cause.

Development and validation of a nomogram for predicting 28-day mortality in patients with ischemic stroke

BACKGROUND/AIM

We aimed to construct a validated nomogram model for predicting short-term (28-day) ischemic stroke mortality among critically ill populations.

MATERIALS AND METHODS

We collected raw data from the Medical Information Mart for Intensive Care IV database, a comprehensive repository renowned for its depth and breadth in critical care information. Subsequently, a rigorous analytical framework was employed, incorporating a 10-fold cross-validation procedure to ensure robustness and reliability. Leveraging advanced statistical methodologies, specifically the least absolute shrinkage and selection operator regression, variables pertinent to 28-day mortality in ischemic stroke were meticulously screened. Next, binary logistic regression was utilized to establish nomogram, then applied concordance index to evaluate discrimination of the prediction models. Predictive performance of the nomogram was assessed by integrated discrimination improvement (IDI) and net reclassification index (NRI). Additionally, we generated calibration curves to assess calibrating ability. Finally, we evaluated the nomogram's net clinical benefit using decision curve analysis (DCA), in comparison with scoring systems clinically applied under common conditions.

RESULTS

A total of 2089 individuals were identified and assigned into training (n = 1443) or validation (n = 646) cohorts. Various identified risk factors, including age, ethnicity, marital status, underlying metastatic solid tumor, Charlson comorbidity index, heart rate, Glasgow coma scale, glucose concentrations, white blood cells, sodium concentrations, potassium concentrations, mechanical ventilation, use of heparin and mannitol, were associated with short-term (28-day) mortality in ischemic stroke individuals. A concordance index of 0.834 was obtained in the training dataset, indicating that our nomogram had good discriminating ability. Results of IDI and NRI in both cohorts proved that our nomogram had positive improvement of predictive performance, compared to other scoring systems. The actual and predicted incidence of mortality showed favorable concordance on calibration curves (P > 0.05). DCA curves revealed that, compared with scoring systems clinically used under common conditions, the constructed nomogram yielded a greater net clinical benefit.

CONCLUSIONS

Utilizing a comprehensive array of fourteen readily accessible variables, a prognostic nomogram was meticulously formulated and rigorously validated to provide precise prognostication of short-term mortality within the ischemic stroke cohort.

Clinical nomogram prediction model to assess the risk of prolonged ICU length of stay in patients with diabetic ketoacidosis: a retrospective analysis based on the MIMIC-IV database

BACKGROUND

The duration of hospitalization, especially in the intensive care unit (ICU), for patients with diabetic ketoacidosis (DKA) is influenced by patient prognosis and treatment costs. Reducing ICU length of stay (LOS) in patients with DKA is crucial for optimising healthcare resources utilization. This study aimed to establish a nomogram prediction model to identify the risk factors influencing prolonged LOS in ICU-managed patients with DKA, which will serve as a basis for clinical treatment, healthcare safety, and quality management research.

METHODS

In this single-centre retrospective cohort study, we performed a retrospective analysis using relevant data extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Clinical data from 669 patients with DKA requiring ICU treatment were included. Variables were selected using the Least Absolute Shrinkage and Selection Operator (LASSO) binary logistic regression model. Subsequently, the selected variables were subjected to a multifactorial logistic regression analysis to determine independent risk factors for prolonged ICU LOS in patients with DKA. A nomogram prediction model was constructed based on the identified predictors. The multivariate variables included in this nomogram prediction model were the Oxford acute severity of illness score (OASIS), Glasgow coma scale (GCS), acute kidney injury (AKI) stage, vasoactive agents, and myocardial infarction.

RESULTS

The prediction model had a high predictive efficacy, with an area under the curve value of 0.870 (95% confidence interval [CI], 0.831-0.908) in the training cohort and 0.858 (95% CI, 0.799-0.916) in the validation cohort. A highly accurate predictive model was depicted in both cohorts using the Hosmer-Lemeshow (H-L) test and calibration plots.

CONCLUSION

The nomogram prediction model proposed in this study has a high clinical application value for predicting prolonged ICU LOS in patients with DKA. This model can help clinicians identify patients with DKA at risk of prolonged ICU LOS, thereby enhancing prompt intervention and improving prognosis.

The value of five scoring systems in predicting the prognosis of patients with sepsis-associated acute respiratory failure

Our study aimed to identify the optimal scoring system for predicting the prognosis of patients with sepsis-associated acute respiratory failure (SA-ARF). All data were taken from the fourth version of the Markets in Intensive Care Medicine (MIMIC-IV) database. Independent risk factors for death in hospitals were confirmed by regression analysis. The predictive value of the five scoring systems was evaluated by receiving operating characteristic (ROC) curves. Kaplan‒Meier curves showed the impact of acute physiology score III (APSIII) on survival and prognosis in patients with SA-ARF. Decision curve analysis (DCA) identified a scoring system with the highest net clinical benefit. ROC curve analysis showed that APS III (AUC: 0.755, 95% Cl 0.714-0.768) and Logical Organ Dysfunction System (LODS) (AUC: 0.731, 95% Cl 0.717-0.7745) were better than Simplified Acute Physiology Score II (SAPS II) (AUC: 0.727, 95% CI 0.713-0.741), Oxford Acute Severity of Illness Score (OASIS) (AUC: 0.706, 95% CI 0.691-0.720) and Sequential Organ Failure Assessment (SOFA) (AUC: 0.606, 95% CI 0.590-0.621) in assessing in-hospital mortality. Kaplan‒Meier survival analysis patients in the high-APS III score group had a considerably poorer median survival time. The DCA curve showed that APS III may provide better clinical benefits for patients. We demonstrated that the APS III score is an excellent predictor of in-hospital mortality.

Interpretable machine learning models for predicting 90-day death in patients in the intensive care unit with epilepsy

PURPOSE

This study aims to develop a machine learning-based model for predicting mortality risk in patients with epilepsy admitted to the intensive care unit (ICU), providing clinicians with an accurate prognostic tool to guide individualized treatment.

METHODS

We collected clinical data from clinical databases (MIMIC IV and eICU-CRD) of epilepsy patients 24 h after ICU admission. The clinical characteristics of ICU patients with epilepsy were carefully feature selected and processed. MIMIC IV as the training set and eICU-CRD database as the test set. Six models were developed and validated, and the best LightGBM model was selected by performance comparison and analysed for interpretability.

RESULTS

The final cohort comprised 429 patients for training and 1217 for testing. The training set exhibited a 90-day mortality rate of 9.32 %, and the test set had an in-hospital 90-day mortality rate of 4.10 %. Utilizing the LightGBM model, we achieved an AUC of 0.956 in the training set. External validation demonstrated promising results with accuracy of 0.898, precision of 0.975, AUC of 0.781, F1 score of 0.945, highlighting the model's potential for guiding clinical decision-making. Significant factors influencing model performance included the severity of illness, as measured by the OASIS score, and clinical parameters like heart rate and body temperature.

CONCLUSION

This study introduces a machine learning-based approach to predict mortality risk in ICU epilepsy patients, offering a valuable tool for clinicians to identify high-risk individuals and devise personalized treatment strategies, thus improving patient prognosis and treatment outcomes.

Analysis of pathogenic factors on the death rate of sepsis patients

BACKGROUND

The Surviving Sepsis Campaign (SSC) believed that early identification of septic shock, aggressive fluid resuscitation and maintenance of effective perfusion pressure should be carried out. However, some of the current research focused on a single death factor for sepsis patients, based on a limited sample, and the research results of the relationship between comorbidities and sepsis related death also have some controversies.

METHOD

Therefore, our study used data from a large sample of 9,544 sepsis patients aged 18-85 obtained from the MIMIC-IV database, to explore the risk factors of death in patients with sepsis. We used the general clinical information, organ dysfunction scores, and comorbidities to analyze the independent risk factors for death of these patients.

RESULTS

The death group had significantly higher organ dysfunction scores, lower BMI, lower body temperature, faster heart rate and lower urine-output. Among the comorbidities, patients suffering from congestive heart failure and liver disease had a higher mortality rate.

CONCLUSION

This study helps to identify sepsis early, based on a comprehensive evaluation of a patient's basic information, organ dysfunction scores and comorbidities, and this methodology could be used for actual clinical diagnosis in hospitals.

A comparison of random survival forest and Cox regression for prediction of mortality in patients with hemorrhagic stroke

OBJECTIVE

To evaluate RSF and Cox models for mortality prediction of hemorrhagic stroke (HS) patients in intensive care unit (ICU).

METHODS

In the training set, the optimal models were selected using five-fold cross-validation and grid search method. In the test set, the bootstrap method was used to validate. The area under the curve(AUC) was used for discrimination, Brier Score (BS) was used for calibration, positive predictive value(PPV), negative predictive value(NPV), and F1 score were combined to compare.

RESULTS

A total of 2,990 HS patients were included. For predicting the 7-day mortality, the mean AUCs for RSF and Cox regression were 0.875 and 0.761, while the mean BS were 0.083 and 0.108. For predicting the 28-day mortality, the mean AUCs for RSF and Cox regression were 0.794 and 0.649, while the mean BS were 0.129 and 0.174. The mean AUCs of RSF and Cox versus conventional scores for predicting patients' 7-day mortality were 0.875 (RSF), 0.761 (COX), 0.736 (SAPS II), 0.723 (OASIS), 0.632 (SIRS), and 0.596 (SOFA), respectively.

CONCLUSIONS

RSF provided a better clinical reference than Cox. Creatine, temperature, anion gap and sodium were important variables in both models.

Evaluating equitable care in the ICU: Creating a causal inference framework to assess the impact of life-sustaining interventions across racial and ethnic groups

Background

Variability in the provision of intensive care unit (ICU)-interventions may lead to disparities between socially defined racial-ethnic groups.

Research Question

We used causal inference to examine the use of invasive mechanical ventilation (IMV), renal replacement therapy (RRT), and vasopressor agents (VP) to identify disparities in outcomes across race-ethnicity in patients with sepsis.

Study Design and Methods

Single-center, academic referral hospital in Boston, Massachusetts, USA. Retrospective analysis of treatment effect with a targeted trial design categorized by treatment assignment within the first 24 hours in the MIMIC-IV dataset (2008- 2019) using targeted maximum likelihood estimation. Of 76,943 ICU stays in MIMIC-IV, 32,971 adult stays fulfilling sepsis-3 criteria were included. The primary outcome was in-hospital mortality. Secondary outcomes were hospital-free days, and occurrence of nosocomial infection stratified by predicted mortality probability ranges and self-reported race-ethnicity. Average treatment effects by treatment type and race-ethnicity, Racial-ethnic group (REG) or White group (WG), were estimated.

Results

Of 19,419 admissions that met inclusion criteria, median age was 68 years, 57.4% were women, 82% were White, and mortality was 18.2%. There was no difference in mortality benefit associated with the administration of IMV, RRT, or VP between the REG and the WG. There was also no difference in hospital-free days or nosocomial infections. These findings are unchanged with different eligibility periods.

Interpretation

There were no differences in the treatment outcomes from three life-sustaining interventions in the ICU according to race-ethnicity. While there was no discernable harm from the treatments across mortality risk, there was also no measurable benefit. These findings highlight the need for research to understand better the risk-benefit of life-sustaining interventions in the ICU.

Illness severity assessment of older adults in critical illness using machine learning (ELDER-ICU): an international multicentre study with subgroup bias evaluation

BACKGROUND

Comorbidity, frailty, and decreased cognitive function lead to a higher risk of death in elderly patients (more than 65 years of age) during acute medical events. Early and accurate illness severity assessment can support appropriate decision making for clinicians caring for these patients. We aimed to develop ELDER-ICU, a machine learning model to assess the illness severity of older adults admitted to the intensive care unit (ICU) with cohort-specific calibration and evaluation for potential model bias.

METHODS

In this retrospective, international multicentre study, the ELDER-ICU model was developed using data from 14 US hospitals, and validated in 171 hospitals from the USA and Netherlands. Data were extracted from the Medical Information Mart for Intensive Care database, electronic ICU Collaborative Research Database, and Amsterdam University Medical Centers Database. We used six categories of data as predictors, including demographics and comorbidities, physical frailty, laboratory tests, vital signs, treatments, and urine output. Patient data from the first day of ICU stay were used to predict in-hospital mortality. We used the eXtreme Gradient Boosting algorithm (XGBoost) to develop models and the SHapley Additive exPlanations method to explain model prediction. The trained model was calibrated before internal, external, and temporal validation. The final XGBoost model was compared against three other machine learning algorithms and five clinical scores. We performed subgroup analysis based on age, sex, and race. We assessed the discrimination and calibration of models using the area under receiver operating characteristic (AUROC) and standardised mortality ratio (SMR) with 95% CIs.

FINDINGS

Using the development dataset (n=50 366) and predictive model building process, the XGBoost algorithm performed the best in all types of validations compared with other machine learning algorithms and clinical scores (internal validation with 5037 patients from 14 US hospitals, AUROC=0·866 [95% CI 0·851-0·880]; external validation in the US population with 20 541 patients from 169 hospitals, AUROC=0·838 [0·829-0·847]; external validation in European population with 2411 patients from one hospital, AUROC=0·833 [0·812-0·853]; temporal validation with 4311 patients from one hospital, AUROC=0·884 [0·869-0·897]). In the external validation set (US population), the median AUROCs of bias evaluations covering eight subgroups were above 0·81, and the overall SMR was 0·99 (0·96-1·03). The top ten risk predictors were the minimum Glasgow Coma Scale score, total urine output, average respiratory rate, mechanical ventilation use, best state of activity, Charlson Comorbidity Index score, geriatric nutritional risk index, code status, age, and maximum blood urea nitrogen. A simplified model containing only the top 20 features (ELDER-ICU-20) had similar predictive performance to the full model.

INTERPRETATION

The ELDER-ICU model reliably predicts the risk of in-hospital mortality using routinely collected clinical features. The predictions could inform clinicians about patients who are at elevated risk of deterioration. Prospective validation of this model in clinical practice and a process for continuous performance monitoring and model recalibration are needed.

FUNDING

National Institutes of Health, National Natural Science Foundation of China, National Special Health Science Program, Health Science and Technology Plan of Zhejiang Province, Fundamental Research Funds for the Central Universities, Drug Clinical Evaluate Research of Chinese Pharmaceutical Association, and National Key R&D Program of China.

Real-world data analysis for the association of glucose control and mortality in critically ill patients

Existing results regarding the usage of glycemic control in critically ill patients for reduced morbidity and mortality have been based on clinical studies but could not be reproduced in large prospective studies. Current guidelines for glycemic control suggest a target blood glucose of 140-180 mg/dL, with lower targets being appropriate for some patients. The current study aims to provide additional evidence to this area, through the usage of real-world retrospective data of everyday clinical practice. We have used the large, credentialed access database MIMIC-IV to assess the effect of glycemic control to patient mortality. Glycemic control has been characterized by the percentage of time that the glucose measurements fall within pre-specified glucose bands. Results from logistic regression and survival analysis are reported, along with visualizations based on methods from the machine learning literature, which all suggest that increased time in low and high glucose values is related to increased ICU mortality and decreased survival.

Predicting unplanned readmissions in the intensive care unit: a multimodality evaluation

A hospital readmission is when a patient who was discharged from the hospital is admitted again for the same or related care within a certain period. Hospital readmissions are a significant problem in the healthcare domain, as they lead to increased hospitalization costs, decreased patient satisfaction, and increased risk of adverse outcomes such as infections, medication errors, and even death. The problem of hospital readmissions is particularly acute in intensive care units (ICUs), due to the severity of the patients' conditions, and the substantial risk of complications. Predicting Unplanned Readmissions in ICUs is a challenging task, as it involves analyzing different data modalities, such as static data, unstructured free text, sequences of diagnoses and procedures, and multivariate time-series. Here, we investigate the effectiveness of each data modality separately, then alongside with others, using state-of-the-art machine learning approaches in time-series analysis and natural language processing. Using our evaluation process, we are able to determine the contribution of each data modality, and for the first time in the context of readmission, establish a hierarchy of their predictive value. Additionally, we demonstrate the impact of Temporal Abstractions in enhancing the performance of time-series approaches to readmission prediction. Due to conflicting definitions in the literature, we also provide a clear definition of the term Unplanned Readmission to enhance reproducibility and consistency of future research and to prevent any potential misunderstandings that could result from diverse interpretations of the term. Our experimental results on a large benchmark clinical data set show that Discharge Notes written by physicians, have better capabilities for readmission prediction than all other modalities.

Outcome prediction for critical care patients with respiratory neoplasms using a multilayer perceptron neural network

OBJECTIVE

The variation in mortality rates of intensive care unit oncological patients may imply that clinical characteristics and prognoses are very different between specific subsets of patients with cancer. The specific characteristics of patients with cancer have not been included as risk factors in the established severity-of-illness scoring systems and comorbidity scores, showing limitations in predicting mortality risk. This study aimed to devise a predictive tool for in-hospital mortality for adult patients with a respiratory neoplasm admitted to the intensive care unit, using an artificial neural network.

METHODS

A total of 1,221 stays in the intensive care unit from the Beth Israel Deaconess Medical Center were studied. The primary endpoint was the all-cause in-hospital mortality prediction. An artificial neural network was developed and compared with six severity-of-illness scores and one comorbidity score. Model building was based on important predictors of lung cancer mortality, such as several laboratory parameters, demographic parameters, organ-supporting treatments, and other clinical information. Discrimination and calibration were assessed.

RESULTS

The AUROC for the multilayer perceptron was 0.885, while it was <0.74 for the conventional systems. The AUPRC for the multilayer perceptron was 0.731, whereas it was ≤0.482 for the conventional systems. The superiority of multilayer perceptron was statistically significant for all pairwise AUROC and AUPRC comparisons. The Brier Score was better for the multilayer perceptron (0.109) than for OASIS (0.148), SAPS III (0.163), and SAPS II (0.154).

CONCLUSION

Discrimination was excellent for multilayer perceptron, which may be a valuable tool for assessing critically ill patients with lung cancer.

Predictive value of the Oxford Acute Severity of Illness Score in acute stroke patients with stroke-associated pneumonia

Background

Stroke-associated pneumonia (SAP) is associated with a poor prognosis and a high mortality rate in stroke patients. However, the accuracy of early prediction of SAP is insufficient, and there is a lack of effective prognostic evaluation methods. Therefore, in this study, we investigated the predictive value of the Oxford Acute Severity of Illness Score (OASIS) in SAP to provide a potential reference index for the incidence and prognosis of SAP.

Methods

We recruited a total of 280 patients with acute ischemic stroke who had been diagnosed and treated in the Zhumadian Central Hospital between January 2021 and January 2023. These patients were divided into an SAP group (86 cases) and a non-SAP group (194 cases) according to SAP diagnostic criteria by expert consensus on the diagnosis and treatment of SAP. We collated general and clinical data from all patients, including the survival of SAP patients during the follow-up period. Multivariate logistic regression was used to analyze the risk factors for SAP. Kaplan-Meier and multivariate COX regression analyses were used to investigate the relationship between OASIS and the prognosis of SAP, and a receiver operating characteristic (ROC) curve was drawn to analyze the predictive value of OASIS for SAP.

Results

Our analyses identified body temperature, C-reactive protein, procalcitonin, OASIS, and a prolonged length of intensive care unit (ICU) stay as the main risk factors for SAP (all Ps < 0.05). Advanced age and an elevated OASIS were identified as the main risk factors for death in SAP patients (all Ps < 0.05). The risk of death in patients with OASIS of 31-42 points was significantly higher than that in patients with OASIS of 12-20 points (HR = 5.588, 95% CI = 1.531-20.401, P = 0.009). ROC curve analysis further showed that OASIS had a high predictive value for morbidity and the incidence of death in SAP patients.

Conclusion

OASIS can effectively predict the onset and death of SAP patients and provides a potential reference index for early diagnosis and the prediction of prognosis in patients with SAP. Our findings should be considered in clinical practice.

Development and verification of a nomogram for predicting short-term mortality in elderly ischemic stroke populations

Stroke is a major healthcare problem worldwide, particularly in the elderly population. Despite limited research on the development of prediction models for mortality in elderly individuals with ischemic stroke, our study aimed to address this knowledge gap. By leveraging data from the Medical Information Mart for Intensive Care IV database, we collected comprehensive raw data pertaining to elderly patients diagnosed with ischemic stroke. Through meticulous screening of clinical variables associated with 28-day mortality, we successfully established a robust nomogram. To assess the performance and clinical utility of our nomogram, various statistical analyses were conducted, including the concordance index, integrated discrimination improvement (IDI), net reclassification index (NRI), calibration curves and decision curve analysis (DCA). Our study comprised a total of 1259 individuals, who were further divided into training (n = 894) and validation (n = 365) cohorts. By identifying several common clinical features, we developed a nomogram that exhibited a concordance index of 0.809 in the training dataset. Notably, our findings demonstrated positive improvements in predictive performance through the IDI and NRI analyses in both cohorts. Furthermore, calibration curves indicated favorable agreement between the predicted and actual incidence of mortality (P > 0.05). DCA curves highlighted the substantial net clinical benefit of our nomogram compared to existing scoring systems used in routine clinical practice. In conclusion, our study successfully constructed and validated a prognostic nomogram, which enables accurate short-term mortality prediction in elderly individuals with ischemic stroke.

Triglyceride-glucose index is associated with the occurrence and prognosis of cardiac arrest: a multicenter retrospective observational study

BACKGROUND

Triglyceride-glucose (TyG) index is an efficient indicator of insulin resistance and is proven to be a valuable marker in several cardiovascular diseases. However, the relationship between TyG index and cardiac arrest (CA) remains unclear. The present study aimed to investigate the association of the TyG index with the occurrence and clinical outcomes of CA.

METHODS

In this retrospective, multicenter, observational study, critically ill patients, including patients post-CA, were identified from the eICU Collaborative Research Database and evaluated. The TyG index for each patient was calculated using values of triglycerides and glucose recorded within 24 h of intensive care unit (ICU) admission. In-hospital mortality and ICU mortality were the primary clinical outcomes. Logistic regression, restricted cubic spline (RCS), and correlation analyses were performed to explore the relationship between the TyG index and clinical outcomes. Propensity score matching (PSM), overlap weighting (OW), and inverse probability of treatment weighting (IPTW) were adopted to balance the baseline characteristics of patients and minimize selection bias to confirm the robustness of the results. Subgroup analysis based on different modifiers was also performed.

RESULTS

Overall, 24,689 critically ill patients, including 1021 patients post-CA, were enrolled. The TyG index was significantly higher in patients post-CA than in those without CA (9.20 (8.72-9.69) vs. 8.89 (8.45-9.41)), and the TyG index had a moderate discrimination ability to identify patients with CA from the overall population (area under the curve = 0.625). Multivariate logistic regression indicated that the TyG index was an independent risk factor for in-hospital mortality (OR = 1.28, 95% CI: 1.03-1.58) and ICU mortality (OR = 1.27, 95% CI: 1.02-1.58) in patients post-CA. RCS curves revealed that an increased TyG index was linearly related to higher risks of in-hospital and ICU mortality (P for nonlinear: 0.225 and 0.271, respectively). Even after adjusting by PSM, IPTW, and OW, the TyG index remained a risk factor for in-hospital mortality and ICU mortality in patients experiencing CA, which was independent of age, BMI, sex, etc. Correlation analyses revealed that TyG index was negatively correlated with the neurological status of patients post-CA.

CONCLUSION

Elevated TyG index is significantly associated with the occurrence of CA and higher mortality risk in patients post-CA. Our findings extend the landscape of TyG index in cardiovascular diseases, which requires further prospective cohort study.

Association of dynamic change of triglyceride-glucose index during hospital stay with all-cause mortality in critically ill patients: a retrospective cohort study from MIMIC IV2.0

BACKGROUND

Biomarker of insulin resistance, namely triglyceride-glucose index, is potentially useful in identifying critically ill patients at high risk of hospital death. However, the TyG index might have variations over time during ICU stay. Hence, the purpose of the current research was to verify the associations between the dynamic change of the TyG index during the hospital stay and all-cause mortality.

METHODS

The present retrospective cohort study was conducted using the Medical Information Mart for Intensive Care IV 2.0 (MIMIC-IV) critical care dataset, which included data from 8835 patients with 13,674 TyG measurements. The primary endpoint was 1-year all-cause mortality. Secondary outcomes included in-hospital all-cause mortality, the need for mechanical ventilation during hospitalization, length of stay in the hospital. Cumulative curves were calculated using the Kaplan-Meier method. Propensity score matching was performed to reduce any potential baseline bias. Restricted cubic spline analysis was also employed to assess any potential non-linear associations. Cox proportional hazards analyses were performed to examine the association between the dynamic change of TyG index and mortality.

RESULTS

The follow-up period identified a total of 3010 all-cause deaths (35.87%), of which 2477 (29.52%) occurred within the first year. The cumulative incidence of all-cause death increased with a higher quartile of the TyGVR, while there were no differences in the TyG index. Restricted cubic spline analysis revealed a nearly linear association between TyGVR and the risk of in-hospital all-cause mortality (P for non-linear = 0.449, P for overall = 0.004) as well as 1-year all-cause mortality (P for non-linear = 0.909, P for overall = 0.019). The area under the curve of all-cause mortality by various conventional severity of illness scores significantly improved with the addition of the TyG index and TyGVR. The results were basically consistent in subgroup analysis.

CONCLUSIONS

Dynamic change of TyG during hospital stay is associated with in-hospital and 1-year all-cause mortality, and may be superior to the effect of baseline TyG index.

Urine output for predicting in-hospital mortality of intensive care patients with cardiogenic shock

Background

The role of urine output (UO) in the first 24 h of admission in the clinical management of cardiogenic shock (CS) patients has not been elucidated.

Methods

This study retrospectively analyzed intensive care CS patients in the MIMIC-IV database. Binomial logistic regression analysis was conducted to evaluate whether UO was an independent risk factor for in-hospital mortality in CS patients. The performance of UO in predicting mortality was evaluated by the receiver operating characteristic (ROC) curve and compared with the Oxford Acute Severity of Illness Score (OASIS). The clinical net benefit of UO in predicting mortality was determined using the decision curve analysis (DCA). Survival analysis was performed with Kaplan-Meier curves.

Results

After adjusting for confounding factors including diuretic use and acute kidney injury (AKI), UO remained an independent risk factor for in-hospital mortality in CS patients. The areas under the ROC curves (AUCs) of UO for predicting in-hospital mortality were 0.712 (UO, ml/day) and 0.701 (UO, ml/kg/h), which were comparable to OASIS (AUC = 0.695). In terms of clinical net benefit, UO was comparable to OASIS, with different degrees of benefit at different threshold probabilities. Survival analysis showed that the risk of in-hospital death in the low-UO (≤857 ml/day) group was 3.0143 times that of the high-UO (>857 ml/day) group.

Conclusions

UO in the first 24 h of admission is an independent risk factor for in-hospital mortality in intensive care CS patients and has moderate predictive value in predicting in-hospital mortality.

Restricted Polypharmacy Compared to Usual Care in Mechanically Ventilated Patients: A Retrospective Cohort Study

BACKGROUND

Adverse effects of excessive sedation in critically ill mechanically ventilated patients are well described. Although guidelines strongly recommend minimizing sedative use, additional agents are added as infusions, often empirically. The tradeoffs associated with such decisions remain unclear.

METHODS

To test the hypothesis that a pragmatic propofol-based sedation regimen with restricted polypharmacy (RP; ie, prohibits additional infusions unless a predefined propofol dosage threshold is exceeded) would increase coma-and ventilator-free days compared with usual care (UC), we performed a retrospective cohort study of adults admitted to intensive care units (ICUs) of a tertiary-level medical center who were mechanically ventilated, initiated on propofol infusion, and had >50% probability of need for continued ventilation for the next 24 hours. We compared RP to UC, adjusting for baseline and time-varying confounding (demographics, care unit, calendar time of admission, vitals, laboratories, other interventions such as vasopressors and fluids, and more) through inverse probability weighting in a target trial framework. Ventilator-free days and coma-free days within 30 days of intubation and in-hospital mortality were the outcomes of interest.

RESULTS

A total of 7974 patients were included in the analysis, of which 3765 followed the RP strategy until extubation. In the full cohort under UC, mean coma-free days were 23.5 (95% confidence interval [CI], [23.3-23.7]), mean ventilator-free days were 20.6 (95% CI, [20.4-20.8]), and the in-hospital mortality rate was 22.0% (95% CI, [21.2-22.8]). We estimated that an RP strategy would increase mean coma-free days by 1.0 days (95% CI, [0.7-1.3]) and ventilator-free days by 1.0 days (95% CI, [0.7-1.3]) relative to UC in our cohort. Our estimate of the confounding-adjusted association between RP and in-hospital mortality was uninformative (-0.5%; 95% CI, [-3.0 to 1.9]).

CONCLUSIONS

Compared with UC, RP was associated with more coma- and ventilator-free days. Restricting addition of adjunct infusions to propofol may represent a viable strategy to reduce duration of coma and mechanical ventilation. These hypothesis-generating findings should be confirmed in a randomized control trial.

Effect of Admission and Discharge Times on Hospital Mortality in Patients With Sepsis

OBJECTIVES

To assess whether the time of admission/discharge time from the ICU and weekend admission are independently associated with hospital mortality in critically ill patients with sepsis.

DESIGN

Retrospective study. Each 24-hour period (08:00 to 07:59 hr) was split into three time periods, defined as "day" (08:00 to 16:59 hr), "evening" (17:00 to 23:59 hr), and "night" (00:00 to 07:59 hr). Weekends were defined as 17:00 hours on Friday to 07:59 hours on Monday. Multivariate logistic regression models were conducted to assess the association between the ICU admission/discharge time, weekend admission, and hospital mortality.

SETTING

Single-center ICUs in China.

PATIENTS

Characteristics and clinical outcomes of 1,341 consecutive septic patients admitted to the emergency ICU, general ICU, or cardiovascular ICU in a tertiary teaching hospital were collected.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

ICU mortality rates were 5.8%, 11.9%, and 10.6%, and hospital mortality rates were 7.3%, 15.6%, and 17.1% during the day, evening, and night time, respectively. Hospital mortality was adjusted for patient to nurse (P/N) ratio, disease severity, Charlson index, age, gender, mechanical ventilation, and shock. Notably, ICU admission time and weekend admission were not predictors of mortality after adjustment. The P/N ratio at admission was significantly associated with mortality ( p < 0.05). The P/N ratio and compliance with the Surviving Sepsis Campaign (SSC) were significantly correlated. After risk adjustment for illness severity at time of ICU discharge and Charlson index, the time of discharge was no longer a significant predictor of mortality.

CONCLUSIONS

ICU admission/discharge time and weekend admission were not independent risk factors of hospital mortality in critically ill patients with sepsis. The P/N ratio at admission, which can affect the compliance rate with SSC, was a predictor of hospital survival. Unstable state on transfer from the ICU was the main risk factor for in-hospital death. These findings may have implications for the management of septic patients.

Association between histamine 2 receptor antagonists and sepsis outcomes in ICU patients: a retrospective analysis using the MIMI-IV database

Background

Sepsis is marked by elevated histamine, which is a vasodilator that increases vascular permeability. Although human studies are lacking, murine models of sepsis have indicated potential protective effects of histamine 2 receptor antagonist administration (H2RAs).

Objective

To assess any association between H2RA use in sepsis-3 patients admitted to the ICU and mortality, mechanical ventilation, length of stay, and markers of renal, liver, and lung dysfunction.

Design

Retrospective cohort study.

Setting

Intensive care units of the Beth Israel Deaconess Medical Center (BIDMC) accessed via the MIMIC-IV database spanning an 11-year period from 2008 to 2019.

Patients (or participants)

A total of 30,591 patients met the inclusion criteria for sepsis-3 on admission (mean age 66.49, standard deviation 15.92).

Main measures

We collected patient age, gender, ethnicity, comorbidities (contained within the Charlson comorbidity index), SOFA score, OASIS score, APS III score, SAPS II score, H2RA use, creatinine, BUN, ALT, AST, and P/F ratios. Primary outcomes were mortality, mechanical ventilation, and ICU length of stay.

Key results

A total of 30,591 patients met inclusion criteria over the 11-year sample period. The 28-day in hospital mortality rate was significantly lower among patients who received an H2RA (12.6% vs 15.1%, p < 0.001) as compared to those who did not receive an H2RA. Patients receiving an H2RA had significantly lower adjusted odds of mortality (0.802, 95% CI 0.741–0.869, p < 0.001), but significantly higher adjusted odds of invasive mechanical ventilation (4.426, 95% CI 4.132–4.741, p < 0.001) and significantly higher ICU LOS (3.2 days vs. 2.4 days, p < 0.001) as compared to the non-H2RA group. H2RA use was also associated with decreased severity of acute respiratory distress syndrome (ARDS) and lower serum creatinine.

Conclusion

Among patients hospitalized in the ICU for sepsis, the use of an H2RA was associated with significantly lower odds of mortality, decreased severity of ARDS, and a lower incidence of renal insufficiency.

Developing an Interpretable Machine Learning Model to Predict in-Hospital Mortality in Sepsis Patients: A Retrospective Temporal Validation Study

BACKGROUND

Risk stratification plays an essential role in the decision making for sepsis management, as existing approaches can hardly satisfy the need to assess this heterogeneous population. We aimed to develop and validate a machine learning model to predict in-hospital mortality in critically ill patients with sepsis.

METHODS

Adult patients fulfilling the definition of Sepsis-3 were included at a large tertiary medical center. Relevant clinical features were extracted within the first 24 h in ICU, re-classified into different genres, and utilized for model development under three strategies: "Basic + Lab", "Basic + Intervention", and "Whole" feature sets. Extreme gradient boosting (XGBoost) was compared with logistic regression (LR) and established severity scores. Temporal validation was conducted using admissions from 2017 to 2019.

RESULTS

The final cohort included 24,272 patients, of which 4013 patients formed the test cohort for temporal validation. The trained and fine-tuned XGBoost model with the whole feature set showed the best discriminatory ability in the test cohort with AUROC as 0.85, significantly higher than the XGBoost "Basic + Lab" model (0.83), the LR "Whole" model (0.82), SOFA (0.63), SAPS-II (0.73), and LODS score (0.74). The performance in varying subgroups remained robust, and predictors, such as increased urine output and supplemental oxygen therapy, were crucially correlated with improved survival when interpretability was explored.

CONCLUSIONS

We developed and validated a novel XGBoost-based model and demonstrated significantly improved performance to LR and other scores in predicting the mortality risks of sepsis patients in the hospital using features in the first 24 h.

Relationship between stress hyperglycemia ratio and allcause mortality in critically ill patients: Results from the MIMIC-IV database

BACKGROUND

Stress hyperglycemia ratio (SHR) was developed to reduce the impact of long-term chronic glycemic factors on stress hyperglycemia levels, which have been linked to clinical adverse events. However, the relationship between SHR and the short- and long-term prognoses of intensive care unit (ICU) patients remains unclear.

METHODS

We retrospectively analyzed 3,887 ICU patients (cohort 1) whose initial fasting blood glucose and hemoglobin A1c data within 24 hours of admission were available and 3,636 ICU patients (cohort 2) who were followed-up for 1-year using the Medical Information Mart for Intensive Care IV v2.0 database. Patients were divided into two groups based on the optimal cutoff value of SHR, which was determined using the receiver operating characteristic (ROC) curve.

RESULTS

There were 176 ICU deaths in cohort 1 and 378 patients experienced all-cause mortality during 1 year of follow-up in cohort 2. The results of logistic regression revealed that SHR was associated with ICU death (odds ratio 2.92 [95% confidence interval 2.14-3.97] P < 0.001), and non-diabetic patients rather than diabetic patients showed an increased risk of ICU death. As per the Cox proportional hazards model, the high SHR group experienced a higher incidence of 1-year all-cause mortality (hazard ratio 1.55 [95% confidence interval 1.26-1.90] P < 0.001). Moreover, SHR had an incremental effect on various illness scores in predicting ICU all-cause mortality.

CONCLUSION

SHR is linked to ICU death and 1-year all-cause mortality in critically ill patients, and it has an incremental predictive value in different illness scores. Moreover, we found that non-diabetic patients, rather than diabetic patients, showed an increased risk of all-cause mortality.

External Validation of Mortality Prediction Models for Critical Illness Reveals Preserved Discrimination but Poor Calibration

OBJECTIVES

In a recent scoping review, we identified 43 mortality prediction models for critically ill patients. We aimed to assess the performances of these models through external validation.

DESIGN

Multicenter study.

SETTING

External validation of models was performed in the Simple Intensive Care Studies-I (SICS-I) and the Finnish Acute Kidney Injury (FINNAKI) study.

PATIENTS

The SICS-I study consisted of 1,075 patients, and the FINNAKI study consisted of 2,901 critically ill patients.

MEASUREMENTS AND MAIN RESULTS

For each model, we assessed: 1) the original publications for the data needed for model reconstruction, 2) availability of the variables, 3) model performance in two independent cohorts, and 4) the effects of recalibration on model performance. The models were recalibrated using data of the SICS-I and subsequently validated using data of the FINNAKI study. We evaluated overall model performance using various indexes, including the (scaled) Brier score, discrimination (area under the curve of the receiver operating characteristics), calibration (intercepts and slopes), and decision curves. Eleven models (26%) could be externally validated. The Acute Physiology And Chronic Health Evaluation (APACHE) II, APACHE IV, Simplified Acute Physiology Score (SAPS)-Reduced (SAPS-R)' and Simplified Mortality Score for the ICU models showed the best scaled Brier scores of 0.11' 0.10' 0.10' and 0.06' respectively. SAPS II, APACHE II, and APACHE IV discriminated best; overall discrimination of models ranged from area under the curve of the receiver operating characteristics of 0.63 (0.61-0.66) to 0.83 (0.81-0.85). We observed poor calibration in most models, which improved to at least moderate after recalibration of intercepts and slopes. The decision curve showed a positive net benefit in the 0-60% threshold probability range for APACHE IV and SAPS-R.

CONCLUSIONS

In only 11 out of 43 available mortality prediction models, the performance could be studied using two cohorts of critically ill patients. External validation showed that the discriminative ability of APACHE II, APACHE IV, and SAPS II was acceptable to excellent, whereas calibration was poor.

Simplified acute physiology score III is excellent for predicting in-hospital mortality in coronary care unit patients with acute myocardial infarction: A retrospective study

Background

Coronary care unit (CCU) patients with acute myocardial infarction (AMI) lack effective predictors of in-hospital mortality. This study aimed to investigate the performance of four scoring systems in predicting in-hospital mortality in CCU patients with AMI.

Methods

The baseline data, the logistic organ dysfunction system (LODS), the Oxford acute severity of illness score (OASIS), the simplified acute physiology score II (SAPS II), and the simplified acute physiology score III (SAPS III) scores of the patients were extracted from the fourth edition of the Medical Information Mart for Critical Care (MIMIC-IV) database. Independent risk factors for in-hospital mortality were identified by regression analysis. We performed receiver operating characteristic (ROC) curves and compared the area under the curve (AUC) to clarify the predictive value of the four scoring systems. Meanwhile, Kaplan-Meier curves and decision curve analysis (DCA) were performed to determine the optimal scoring system for predicting in-hospital mortality.

Results

A total of 1,098 patients were included. The SAPS III was an independent risk factor for predicting in-hospital mortality in CCU patients with AMI before and after the propensity score matching (PSM) analysis. The discrimination of in-hospital mortality by SAPS III was superior to that of LODS, OASIS, and SAPS II. The AUC of the SAPS III scoring system was the highest among the four scoring systems, at 0.901 (before PSM) and 0.736 (after PSM). Survival analysis showed that significantly more in-hospital mortality occurred in the high-score SAPS III group compared to the low-score SAPS III group before PSM (HR 7.636, P < 0.001) and after PSM (HR 2.077, P = 0.005). The DCA curve of SAPS III had the greatest benefit score across the largest threshold range compared to the other three scoring systems.

Conclusion

The SAPS III was an independent risk factor for predicting in-hospital mortality in CCU patients with AMI. The predictive value for in-hospital mortality with SAPS III is superior to that of LODS, OASIS, and SAPS II. The results of the DCA analysis suggest that SAPS III may provide a better clinical benefit for patients. We demonstrated that SAPS III is an excellent scoring system for predicting in-hospital mortality for CCU patients with AMI.

Red blood cell distribution width at admission predicts outcome in critically ill patients with kidney failure: a retrospective cohort study based on the MIMIC-IV database

PURPOSE

We aimed to explore whether red blood cell distribution width (RDW) could serve as a biomarker to predict outcomes in critically ill patients with kidney failure in this study.

MATERIALS AND METHODS

This retrospective study was conducted with the Medical Information Mart for Intensive Care IV (MIMIC-IV).A total of 674 patients were divided into three groups based on tertiles of RDW. We used the generalized additive model, Kaplan-Meier curve, and Cox proportional hazards models to evaluate the association between RDW and clinical outcomes. We then performed subgroup analyses to investigate the stability of the associations between RDW and all-cause mortality.

RESULTS

Nonlinear and J-shaped curves were observed in the generalized additive model. Kaplan-Meier analysis showed that patients with elevated RDW had a lower survival rate. The Cox regression model indicated that high levels of RDW were most closely associated with ICU mortality and 30-day mortality (HR = 4.71, 95% CI: 1.69-11.64 and HR = 6.62, 95% CI: 2.84-15.41). Subgroup analyses indicated that the associations between RDW and all-cause mortality were stable.

CONCLUSIONS

Elevated levels of RDW were associated with an increased risk of all-cause mortality, and RDW could be an independent prognostic factor for kidney failure.

A cost-aware framework for the development of AI models for healthcare applications

Accurate artificial intelligence (AI) for disease diagnosis could lower healthcare workloads. However, when time or financial resources for gathering input data are limited, as in emergency and critical-care medicine, developing accurate AI models, which typically require inputs for many clinical variables, may be impractical. Here we report a model-agnostic cost-aware AI (CoAI) framework for the development of predictive models that optimize the trade-off between prediction performance and feature cost. By using three datasets, each including thousands of patients, we show that relative to clinical risk scores, CoAI substantially reduces the cost and improves the accuracy of predicting acute traumatic coagulopathy in a pre-hospital setting, mortality in intensive-care patients and mortality in outpatient settings. We also show that CoAI outperforms state-of-the-art cost-aware prediction strategies in terms of predictive performance, model cost, training time and robustness to feature-cost perturbations. CoAI uses axiomatic feature-attribution methods for the estimation of feature importance and decouples feature selection from model training, thus allowing for a faster and more flexible adaptation of AI models to new feature costs and prediction budgets.

Predicting medical events and ICU requirements using a multimodal multiobjective transformer network

Effective utilization of premium hospital resources such as intensive care unit (ICU), operating theater (OT), mechanical ventilator, endotracheal tube, and so on plays a significant role in providing high-quality care to critically ill patients within reasonable costs. Non-availability of specialized resources can lead to dire consequences for such patients, and in the worst case, may even turn out to be fatal. However, these resources cannot be kept idle, as they are expensive to maintain. Therefore, one of the core functions of hospital management is targeted at planning and managing these critical resources in order to provide efficient and effective health-care services to the end-users. Predictive technologies play a big role in this. In this article, we present methods for predicting the length of stay in ICU as well as the need for critical interventions for a patient based on the vital signs, laboratory measurements, and the nursing notes of the patient prepared within the first 24 h of ICU stay. The model has been built and cross-validated on the publicly available Medical Information Mart for Intensive Care (MIMIC-III v1.4) data set. We show that the proposed model performs way better than most of the earlier models in the prediction of ICU stay, which had used patient vitals primarily. Experimental results also demonstrate the advantage of using a multiobjective model over independent models for the prediction of ICU stay and critical interventions. The proposed model uses Local Interpretable Model-agnostic Explanations (LIME) that help in identifying the features responsible for predictive decisions. This is very useful in building trust and confidence in the prediction model among clinical practitioners.

Predicting in-hospital mortality for MIMIC-III patients: A nomogram combined with SOFA score

Predicting the mortality of patients provides a reference for doctors to judge their physical condition. This study aimed to construct a nomogram to improve the prediction accuracy of patients' mortality. Patients with severe diseases were screened from the Medical Information Mart for Intensive Care (MIMIC) III database; 70% of patients were randomly selected as the training set for the model establishment, while 30% were used as the test set. The least absolute shrinkage and selection operator (LASSO) regression method was used to filter variables and select predictors. A multivariable logistic regression fit was used to determine the association between in-hospital mortality and risk factors and to construct a nomogram. A total of 9276 patients were included. The area under the curve (AUC) for the clinical nomogram based on risk factors selected by LASSO and multivariable logistic regressions were 0.849 (95% confidence interval [CI]: 0.835-0.863) and 0.821 (95% CI: 0.795-0.846) in the training and test sets, respectively. Therefore, this nomogram might help predict the in-hospital mortality of patients admitted to the intensive care unit (ICU).

The value of anion gap for predicting the short-term all-cause mortality of critically ill patients with cardiac diseases, based on MIMIC-III database

BACKGROUND

The association of anion gap (AG) with short-term mortality in the critically ill patients with cardiac diseases is still not well understood.

OBJECTIVE

To evaluate the association of AG with short-term mortality, and the predictive ability of AG for short-term mortality in critically ill patients with cardiac diseases.

METHODS

This retrospective cohort study enrolled 9104 critically ill patients with cardiac diseases from the Medical Information Mart for Intensive Care III (MIMIC III) database. The restricted cubic spline models were used to evaluate the nonlinear relationship between AG and short-term mortality. Cox proportional hazards regression models and subgroup analysis were applied to assess the association of AG with short-term mortality.

RESULTS

The data were divided into three groups by AG tertiles: tertile I (AG <12, n = 2095), tertile II (12 ≤ AG < 15, n = 3195), and tertile III (15 ≤ AG, n = 3814). The restricted cubic spline models revealed continuous AG was non-linearly related to short-term mortality. The elevated AG tertiles were strongly associated with higher in-hospital, 30-day and 90-day mortality (all P for trend < 0.001). After adding AG to traditional severity scores, the area under curves (AUCs) elevated significantly compared to severity scores alone (all DeLong's test: P < 0.001). Subgroup analysis did not indicate significant interaction in most diverse subgroups.

CONCLUSION

AG was an independent risk factor for short-term all-cause mortality in critically ill patients with cardiac diseases. AG improved significantly the mortality predictive abilities of traditional severity scores when AG was added to these scores.

Identification of key predictors of hospital mortality in critically ill patients with embolic stroke using machine learning

Embolic stroke (ES) is characterized by high morbidity and mortality. Its mortality predictors remain unclear. The present study aimed to use machine learning (ML) to identify the key predictors of mortality for ES patients in the intensive care unit (ICU). Data were extracted from two large ICU databases: Medical Information Mart for Intensive Care (MIMIC)-IV for training and internal validation, and eICU Collaborative Research Database (eICU-CRD) for external validation. We developed predictive models of ES mortality based on 15 ML algorithms. We relied on the synthetic minority oversampling technique (SMOTE) to address class imbalance. Our main performance metric was area under the receiver operating characteristic (AUROC). We adopted recursive feature elimination (RFE) for feature selection. We assessed model performance using three disease-severity scoring systems as benchmarks. Of the 1566 and 207 ES patients enrolled in the two databases, there were 173 (15.70%), 73 (15.57%), and 36 (17.39%) hospital mortality in the training, internal validation, and external validation cohort, respectively. The random forest (RF) model had the largest AUROC (0.806) in the internal validation phase and was chosen as the best model. The AUROC of the RF compact (RF-COM) model containing the top six features identified by RFE was 0.795. In the external validation phase, the AUROC of the RF model was 0.838, and the RF-COM model was 0.830, outperforming other models. Our findings suggest that the RF model was the best model and the top six predictors of ES hospital mortality were Glasgow Coma Scale, white blood cell, blood urea nitrogen, bicarbonate, age, and mechanical ventilation.

Explainable Machine-Learning Model for Prediction of In-Hospital Mortality in Septic Patients Requiring Intensive Care Unit Readmission

INTRODUCTION

Septic patients requiring intensive care unit (ICU) readmission are at high risk of mortality, but research focusing on the association of ICU readmission due to sepsis and mortality is limited. The aim of this study was to develop and validate a machine learning (ML) model for predicting in-hospital mortality in septic patients readmitted to the ICU using routinely available clinical data.

METHODS

The data used in this study were obtained from the Medical Information Mart for Intensive Care (MIMIC-IV, v1.0) database, between 2008 and 2019. The study cohort included patients with sepsis requiring ICU readmission. The data were randomly split into a training (75%) data set and a validation (25%) data set. Nine popular ML models were developed to predict mortality in septic patients readmitted to the ICU. The model with the best accuracy and area under the curve (A.C.) in the validation cohort was defined as the optimal model and was selected for further prediction studies. The SHAPELY Additive explanations (SHAP) values and Local Interpretable Model-Agnostic Explanation (LIME) methods were used to improve the interpretability of the optimal model.

RESULTS

A total of 1117 septic patients who had required ICU readmission during the study period were enrolled in the study. Of these participants, 434 (38.9%) were female, and the median (interquartile range [IQR]) age was 68.6 (58.4-79.2) years. The median (IQR) ICU interval duration was 2.60 (0.64-5.78) days. After feature selection, 31 of 47 clinical factors were ultimately chosen for use in model construction. Of the nine ML models tested, the best performance was achieved with the random forest (RF) model, with an A.C. of 0.81, an accuracy of 85% and a precision of 62% in the validation cohort. The SHAP summary analysis revealed that Glasgow Coma Scale score, urine output, blood urea nitrogen, lactate, platelet count and systolic blood pressure were the top six most important factors contributing to the RF model. Additionally, the LIME method demonstrated how the RF model works in terms of explaining risk of death prediction in septic patients requiring ICU readmission.

CONCLUSION

The ML models reported here showed a good prognostic prediction ability in septic patients requiring ICU readmission. Of the features selected, the parameters related to organ perfusion made the largest contribution to outcome prediction during ICU readmission in septic patients.

Development and validation of a clinical risk model to predict the hospital mortality in ventilated patients with acute respiratory distress syndrome: a population-based study

Background

Large variability in mortality exists in patients of acute respiratory distress syndrome (ARDS), especially those with invasive ventilation. The aim of this study was to develop a model to predict risk of in-hospital death in ventilated ARDS patients.

Methods

Ventilated patients with ARDS from two public databases (MIMIC-III and eICU-CRD) were randomly divided as training cohort and internal validation cohort. Least absolute shrinkage and selection operator (LASSO) and then Logistic regression was used to construct a predictive model with demographic, clinical, laboratory, comorbidities and ventilation variables ascertained at first 24 h of ICU admission and invasive ventilation. Our model was externally validated using data from another database (MIMIC-IV).

Results

A total of 1075 adult patients from MIMIC-III and eICU were randomly divided into training cohort (70%, n = 752) and internal validation cohort (30%, n = 323). 521 patients were included from MIMIC-IV. From 176 potential predictors, 9 independent predictive factors were included in the final model. Five variables were ascertained within the first 24 h of ICU admission, including age (OR, 1.02; 95% CI: 1.01–1.03), mean of respiratory rate (OR, 1.04; 95% CI: 1.01–1.08), the maximum of INR (OR, 1.14; 95% CI: 1.03–1.31) and alveolo-arterial oxygen difference (OR, 1.002; 95% CI: 1.001–1.003) and the minimum of RDW (OR, 1.17; 95% CI: 1.09–1.27). And four variables were collected within the first 24 h of invasive ventilation: mean of temperature (OR, 0.70; 95% CI: 0.57–0.86), the maximum of lactate (OR, 1.15; 95% CI: 1.09–1.22), the minimum of blood urea nitrogen (OR, 1.02; 95% CI: 1.01–1.03) and white blood cell counts (OR, 1.03; 95% CI: 1.01–1.06). Our model achieved good discrimination (AUC: 0.77, 95% CI: 0.73–0.80) in training cohort but the performance declined in internal (AUC: 0.75, 95% CI: 0.69–0.80) and external validation cohort (0.70, 95% CI: 0.65–0.74) and showed modest calibration.

Conclusions

A risk score based on routinely collected variables at the start of admission to ICU and invasive ventilation can predict mortality of ventilated ARDS patients, with a moderate performance.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-022-02057-0.

Triglyceride-glucose index linked to all-cause mortality in critically ill patients: a cohort of 3026 patients

BACKGROUND

Triglyceride-glucose (TyG) index as a reliable surrogate of insulin resistance (IR) has been shown to be related to adverse clinical outcomes in patients with acute coronary syndrome, heart failure, ischemic stroke and so on. However, the relationship between TyG index and all-cause mortality in intensive care unit (ICU) patients remains unknown. The purpose of this study was to investigate the correlation between TyG index and all-cause mortality to evaluate the impact of IR on the prognosis of this population.

METHODS

This was a retrospective observational study that included 3026 patients who had an initial triglyceride and glucose data on the first day of ICU admission, and all data were extracted from the Medical Information Mart for Intensive Care III (MIMIC-III) database. These patients were grouped into quartiles (Q1-Q4) according to TyG index. The Kaplan-Meier analysis was used to compare all-cause mortality among the above four groups. Cox proportional hazards analyses were performed to examine the association between TyG index and all-cause mortality.

RESULTS

During 10.46 years of follow-up, 1148 (37.9%) patients died, of which 350 (11.6%) occurred during the hospital stay and 258 (8.5%) occurred during the ICU stay. Kaplan-Meier analysis showed that the risk of all-cause mortality was significantly higher in patients with higher TyG index (log-rank P = 0.021). Multivariable Cox proportional hazards analyses showed that the TyG index was an independent risk predictor of ICU death (HR: 1.72, 95% CI 1.18-2.52, P = 0.005) and hospital death (HR: 2.19, 95% CI 1.59-3.03, P < 0.001), and each 1-unit increased in the TyG index, a 1.19-fold increase in the risk of death during the hospital stay.

CONCLUSIONS

TyG index is strongly related to the all-cause mortality increasing in critically ill patients. This finding indicates that the TyG index might be useful in identifying people at high risk of ICU death and hospital death.

Effects of transthoracic echocardiography on the prognosis of patients with acute respiratory distress syndrome: a propensity score matched analysis of the MIMIC-III database

Background

Acute respiratory distress syndrome (ARDS) has high mortality and is mainly related to the circulatory failure.Therefore, real-time monitoring of cardiac function and structural changes has important clinical significance.Transthoracic echocardiography (TTE) is a simple and noninvasive real-time cardiac examination which is widely used in intensive care unit (ICU) patients.The purpose of this study was to analyze the effect of TTE on the prognosis of ICU patients with ARDS.

Methods

The data of ARDS patients were retrieved from the MIMIC-III v1.4 database and patients were divided into the TTE group and non-TTE group. The baseline data were compared between the two groups. The effect of TTE on the prognosis of ARDS patients was analyzed through multivariate logistic analysis and the propensity score (PS). The primary outcome was the 28-d mortality rate. The secondary outcomes included pulmonary artery catheter (PAC) and Pulse index continuous cardiac output (PiCCO) administration, the ventilator-free and vasopressor-free days and total intravenous infusion volume on days 1, 2 and 3 of the mechanical ventilation. To illuminate the effect of echocardiography on the outcomes of ARDS patients,a sensitivity analysis was conducted by excluding those patients receiving either PiCCO or PAC. We also performed a subgroup analysis to assess the impact of TTE timing on the prognosis of patients with ARDS.

Results

A total of 1,346 ARDS patients were enrolled, including 519 (38.6%) cases in the TTE group and 827 (61.4%) cases in the non-TTE group. In the multivariate logistic regression, the 28-day mortality of patients in the TTE group was greatly improved (OR 0.71, 95%CI 0.55–0.92, P = 0.008). More patients in the TTE group received PAC (2% vs. 10%, P < 0.001) and the length of ICU stay in the TTE group was significantly shorter than that in the non-TTE group (17d vs.14d, P = 0.0001). The infusion volume in the TTE group was significantly less than that of the non-TTE group (6.2L vs.5.5L on day 1, P = 0.0012). Importantly, the patients in the TTE group were weaned ventilators earlier than those in the non-TTE group (ventilator-free days within 28 d: 21 d vs. 19.8 d, respectively, P = 0.071). The Kaplan–Meier survival curves showed that TTE patients had significant lower 28-day mortality than non-TTE patients (log-rank = 0.004). Subgroup analysis showed that TTE after hemodynamic disorders can not improve prognosis (OR 1.02, 95%CI 0.79–1.34, P = 0.844).

Conclusion

TTE was associated with improved 28-day outcomes in patients with ARDS.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12890-022-02028-5.

Mortality Prediction in Cardiac Intensive Care Unit Patients: A Systematic Review of Existing and Artificial Intelligence Augmented Approaches

The medical complexity and high acuity of patients in the cardiac intensive care unit make for a unique patient population with high morbidity and mortality. While there are many tools for predictions of mortality in other settings, there is a lack of robust mortality prediction tools for cardiac intensive care unit patients. The ongoing advances in artificial intelligence and machine learning also pose a potential asset to the advancement of mortality prediction. Artificial intelligence algorithms have been developed for application of electrocardiogram interpretation with promising accuracy and clinical application. Additionally, artificial intelligence algorithms applied to electrocardiogram interpretation have been developed to predict various variables such as structural heart disease, left ventricular systolic dysfunction, and atrial fibrillation. These variables can be used and applied to new mortality prediction models that are dynamic with the changes in the patient's clinical course and may lead to more accurate and reliable mortality prediction. The application of artificial intelligence to mortality prediction will fill the gaps left by current mortality prediction tools.

E-CatBoost: An efficient machine learning framework for predicting ICU mortality using the eICU Collaborative Research Database

Improving the Intensive Care Unit (ICU) management network and building cost-effective and well-managed healthcare systems are high priorities for healthcare units. Creating accurate and explainable mortality prediction models helps identify the most critical risk factors in the patients' survival/death status and early detect the most in-need patients. This study proposes a highly accurate and efficient machine learning model for predicting ICU mortality status upon discharge using the information available during the first 24 hours of admission. The most important features in mortality prediction are identified, and the effects of changing each feature on the prediction are studied. We used supervised machine learning models and illness severity scoring systems to benchmark the mortality prediction. We also implemented a combination of SHAP, LIME, partial dependence, and individual conditional expectation plots to explain the predictions made by the best-performing model (CatBoost). We proposed E-CatBoost, an optimized and efficient patient mortality prediction model, which can accurately predict the patients' discharge status using only ten input features. We used eICU-CRD v2.0 to train and validate the models; the dataset contains information on over 200,000 ICU admissions. The patients were divided into twelve disease groups, and models were fitted and tuned for each group. The models' predictive performance was evaluated using the area under a receiver operating curve (AUROC). The AUROC scores were 0.86 [std:0.02] to 0.92 [std:0.02] for CatBoost and 0.83 [std:0.02] to 0.91 [std:0.03] for E-CatBoost models across the defined disease groups; if measured over the entire patient population, their AUROC scores were 7 to 18 and 2 to 12 percent higher than the baseline models, respectively. Based on SHAP explanations, we found age, heart rate, respiratory rate, blood urine nitrogen, and creatinine level as the most critical cross-disease features in mortality predictions.

Interpretability and fairness evaluation of deep learning models on MIMIC-IV dataset

The recent release of large-scale healthcare datasets has greatly propelled the research of data-driven deep learning models for healthcare applications. However, due to the nature of such deep black-boxed models, concerns about interpretability, fairness, and biases in healthcare scenarios where human lives are at stake call for a careful and thorough examination of both datasets and models. In this work, we focus on MIMIC-IV (Medical Information Mart for Intensive Care, version IV), the largest publicly available healthcare dataset, and conduct comprehensive analyses of interpretability as well as dataset representation bias and prediction fairness of deep learning models for in-hospital mortality prediction. First, we analyze the interpretability of deep learning mortality prediction models and observe that (1) the best-performing interpretability method successfully identifies critical features for mortality prediction on various prediction models as well as recognizing new important features that domain knowledge does not consider; (2) prediction models rely on demographic features, raising concerns in fairness. Therefore, we then evaluate the fairness of models and do observe the unfairness: (1) there exists disparate treatment in prescribing mechanical ventilation among patient groups across ethnicity, gender and age; (2) models often rely on racial attributes unequally across subgroups to generate their predictions. We further draw concrete connections between interpretability methods and fairness metrics by showing how feature importance from interpretability methods can be beneficial in quantifying potential disparities in mortality predictors. Our analysis demonstrates that the prediction performance is not the only factor to consider when evaluating models for healthcare applications, since high prediction performance might be the result of unfair utilization of demographic features. Our findings suggest that future research in AI models for healthcare applications can benefit from utilizing the analysis workflow of interpretability and fairness as well as verifying if models achieve superior performance at the cost of introducing bias.

Developing machine learning models for prediction of mortality in the medical intensive care unit

BACKGROUND AND OBJECTIVE

Alert of patient deterioration is essential for prompt medical intervention in the Medical Intensive Care Unit (MICU). Logistic Regression (LR) has been used for the development of most conventional severity-of-illness scoring systems to anticipate the risk of mortality in the MICU. Machine Learning (ML) models such as probabilistic graphical models and Extreme Gradient Boosting (XGB) have demonstrated improved prediction accuracy in patient outcomes compared to LR. The aim was to compare three ML models to the SAPS, SAPS II, SAPS III, SOFA, serial SOFA, LODS, and OASIS for prediction of MICU mortality.

METHODS

A Bayesian Network (BN), Naïve Bayes network (NB), and a XGB model were developed. 9893 adult MICU-stays from the MIMIC-III database were studied. The primary outcome was MICU mortality prediction and the secondary outcome was 1-year mortality prediction. Data analyzed consisted on routine physiological measurements collected during 5 hours in the MICU, demographic and diagnoses/procedure features. The performance was evaluated by accuracy statistics, discrimination and calibration measures. Limitations of the study were discussed.

RESULTS

The AUROC for MICU mortality prediction was 0.919 for XGB, 0.905 for BN, and 0.864 for NB, while the conventional systems displayed much lower values with the serial SOFA having the best value (0.814). The Diagnostic Odds Ratio was ≤7.099 for all the conventional systems, reaching values of 30.115 for XGB and 22.648 for BN. The XGB achieved a sensitivity of 0.831 and specificity of 0.86 assuring an acceptable precision (0.528), whose values were much lower for the conventional systems. The Brier score was better for the ML models, except for the NB (0.119), with 0.072 for XGB and 0.081 for BN.

CONCLUSIONS

The XGB and BN substantially outperformed the conventional systems for discrimination, calibration and the accuracy statistics assessed. The NB showed inferior performance to the XGB and BN but improved the discrimination and all accuracy statistics of the conventional systems except for an inferior calibration and 1-year mortality discrimination. The XGB showed the best performance among all models. These ML models have the potential to improve the monitoring of MICU patients, which must be evaluated in future studies.

A Novel Nomogram for Predicting Morbidity Risk in Patients with Secondary Malignant Neoplasm of Bone and Bone Marrow: An Analysis Based on the Large MIMIC-III Clinical Database

Objective

Bone and bone marrow are the third most frequent sites of metastases from many cancers and are associated with low survival and high morbidity rates. Currently, there are no effective bedside tools to predict the morbidity risk of these patients in general intensive care units (ICUs). The main objective of this study was to establish and validate a nomogram to predict the morbidity risk of patients with bone and bone marrow metastases.

Methods

Data on patients with bone and bone marrow metastases were extracted from the Medical Information Mart for Intensive Care III (MIMIC-III) database. The patients were divided into training and validation cohorts. The data were analyzed using univariate and multivariate Cox regression methods. Factors significantly and independently prognostic of survival were used to construct a nomogram predicting 30-day morbidity. The nomogram was validated by various methods, including Harrell’s concordance index (C-index), area under the receiver operating characteristic curve (AUC), calibration curve, integrated discrimination improvement (IDI), net reclassification index (NRI), and decision curve analysis (DCA).

Results

The study included 610 patients in the training cohort and 262 in the validation cohort. Multivariate Cox regression analysis showed that temperature, SpO₂, Sequential Organ Failure Assessment (SOFA) score, Oxford Acute Severity of Illness Score (OASIS), comorbidities with coagulopathy, white blood cell count, heart rate, and respiratory rate were independent predictors of patient survival. The resulting nomogram had good discriminative ability, as shown by high AUCs, and was well calibrated, as demonstrated by calibration curves. Improvements in NRI and IDI values suggested that the nomogram was superior to the SOFA scoring system. DCA curves revealed that the nomogram showed good value in clinical applications.

Conclusion

This prognostic nomogram, based on demographic and laboratory parameters, was predictive of the 30-day morbidity rate in patients with secondary malignant neoplasms of the bone and bone marrow, suggesting its applicability in clinical practice.

Acute kidney injury-attributable mortality in critically ill patients with sepsis

Background

To assess whether acute kidney injury (AKI) is independently associated with hospital mortality in ICU patients with sepsis, and estimate the excess AKI-related mortality attributable to AKI.

Methods

We analyzed adult patients from two distinct retrospective critically ill cohorts: (1) Medical Information Mart for Intensive Care IV (MIMIC IV; n = 15,610) cohort and (2) Wenzhou (n = 1,341) cohort. AKI was defined by Kidney Disease: Improving Global Outcomes (KDIGO) criteria. We applied multivariate logistic and linear regression models to assess the hospital and ICU mortality, hospital length-of-stay (LOS), and ICU LOS. The excess attributable mortality for AKI in ICU patients with sepsis was further evaluated.

Results

AKI occurred in 5,225 subjects in the MIMIC IV cohort (33.5%) and 494 in the Wenzhou cohort (36.8%). Each stage of AKI was an independent risk factor for hospital mortality in multivariate logistic regression after adjusting for baseline illness severity. The excess attributable mortality for AKI was 58.6% (95% CI [46.8%-70.3%]) in MIMIC IV and 44.6% (95% CI [12.7%-76.4%]) in Wenzhou. Additionally, AKI was independently associated with increased ICU mortality, hospital LOS, and ICU LOS.

Conclusion

Acute kidney injury is an independent risk factor for hospital and ICU mortality, as well as hospital and ICU LOS in critically ill patients with sepsis. Thus, AKI is associated with excess attributable mortality.

The predictive value of the Oxford Acute Severity of Illness Score for clinical outcomes in patients with acute kidney injury

Objective

To compare the performance of the Oxford Acute Severity of Illness Score (OASIS), the Acute Physiology and Chronic Health Evaluation II (APACHE II) score, the Simplified Acute Physiology Score II (SAPS II), and the Sequential Organ Failure Assessment (SOFA) score in predicting 28-day mortality in acute kidney injury (AKI) patients.

Methods

Data were extracted from the Beijing Acute Kidney Injury Trial (BAKIT). A total of 2954 patients with complete clinical data were included in this study. Receiver operating characteristic (ROC) curves were used to analyze and evaluate the predictive effects of the four scoring systems on the 28-day mortality risk of AKI patients and each subgroup. The best cutoff value was identified by the highest combined sensitivity and specificity using Youden’s index.

Results

Among the four scoring systems, the area under the curve (AUC) of OASIS was the highest. The comparison of AUC values of different scoring systems showed that there were no significant differences among OASIS, APACHE II, and SAPS II, which were better than SOFA. Moreover, logistic analysis revealed that OASIS was an independent risk factor for 28-day mortality in AKI patients. OASIS also had good predictive ability for the 28-day mortality of each subgroup of AKI patients.

Conclusion

OASIS, APACHE II, and SAPS II all presented good discrimination and calibration in predicting the 28-day mortality risk of AKI patients. OASIS, APACHE II, and SAPS II had better predictive accuracy than SOFA, but due to the complexity of APACHE II and SAPS II calculations, OASIS is a good substitute.

Trial Registration

This study was registered at www.chictr.org.cn (registration number Chi CTR-ONC-11001875). Registered on 14 December 2011.

Association of Thiamine Use with Outcomes in Patients with Sepsis and Alcohol Use Disorder: An Analysis of the MIMIC-III Database

Introduction

The association between thiamine use and clinical outcomes among patients with sepsis and alcohol use disorder (AUD) is unclear.

Methods

In this retrospective cohort study of patients from Medical Information Mart for Intensive Care III (MIMIC-III, version 1.4), we evaluated the association of thiamine use with clinical outcomes in patients with AUD and sepsis. The primary outcome was 28-day survival, and secondary outcomes included ICU, in-hospital, and 90-day mortality, ICU and hospital length of stay, duration of vasopressor use, need and duration of continuous renal replacement therapy (CRRT), and dynamic changes for variables up to day 7 after ICU admission.

Results

A total of 944 patients with sepsis and AUD were included in this cohort [median age, 53.1 years; women, 26.0% (245 of 944)]. Among all patients, 24.6% (233 of 944) received thiamine with a dose of 200 mg (IQR 100–345 mg). The 28-day mortality was 11.2% (26 of 233) in the thiamine use group compared with 18.6% (132 of 711) in the no thiamine use group (P = 0.009). After adjustment for a series of confounders, the mixed-effects Cox proportional hazards models showed that administration of thiamine was associated with a lower risk of 28-day mortality compared with no administration of thiamine.

Conclusions

In critically ill patients with alcohol use disorder admitted for sepsis, treatment with thiamine may be associated with a decreased risk of death. However, the present results should be interpreted with caution due to the limitations of retrospective design. Additional larger, multicenter randomized controlled trials are needed to confirm our findings.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40121-022-00603-1.

Identification of acute respiratory distress syndrome subphenotypes de novo using routine clinical data: a retrospective analysis of ARDS clinical trials

OBJECTIVES

The acute respiratory distress syndrome (ARDS) is a heterogeneous condition, and identification of subphenotypes may help in better risk stratification. Our study objective is to identify ARDS subphenotypes using new simpler methodology and readily available clinical variables.

SETTING

This is a retrospective Cohort Study of ARDS trials. Data from the US ARDSNet trials and from the international ART trial.

PARTICIPANTS

3763 patients from ARDSNet data sets and 1010 patients from the ART data set.

PRIMARY AND SECONDARY OUTCOME MEASURES

The primary outcome was 60-day or 28-day mortality, depending on what was reported in the original trial. K-means cluster analysis was performed to identify subgroups. Sets of candidate variables were tested to assess their ability to produce different probabilities for mortality in each cluster. Clusters were compared with biomarker data, allowing identification of subphenotypes.

RESULTS

Data from 4773 patients were analysed. Two subphenotypes (A and B) resulted in optimal separation in the final model, which included nine routinely collected clinical variables, namely heart rate, mean arterial pressure, respiratory rate, bilirubin, bicarbonate, creatinine, PaO₂, arterial pH and FiO₂. Participants in subphenotype B showed increased levels of proinflammatory markers, had consistently higher mortality, lower number of ventilator-free days at day 28 and longer duration of ventilation compared with patients in the subphenotype A.

CONCLUSIONS

Routinely available clinical data can successfully identify two distinct subphenotypes in adult ARDS patients. This work may facilitate implementation of precision therapy in ARDS clinical trials.