Mortality prediction using SAPS II: an update for French intensive care units

Evaluation of score-based tertiary triage policies during the COVID-19 pandemic: simulation study with real-world intensive care data

OBJECTIVE

The explicit prohibition of discontinuing intensive care unit (ICU) treatment that has already begun by the newly established German Triage Act in favor of new patients with better prognoses (tertiary triage) under crisis conditions may prevent saving as many patients as possible and therefore may violate the international well-accepted premise of undertaking the "best for the most" patients. During the COVID-19 pandemic, authorities set up lockdown measures and infection-prevention strategies to avoid an overburdened health-care system. In cases of situational overload of ICU resources, when transporting options are exhausted, the question of a tertiary triage of patients arises.

METHODS

We provide data-driven analyses of score- and non-score-based tertiary triage policies using simulation and real-world electronic health record data in a COVID-19 setting. Ten different triage policies, for example, based on the Simplified Acute Physiology Score (SAPS II), are compared based on the resulting mortality in the ICU and inferential statistics.

RESULTS

Our study shows that score-based tertiary triage policies outperform non-score-based tertiary triage policies including compliance with the German Triage Act. Based on our simulation model, a SAPS II score-based tertiary triage policy reduces mortality in the ICU by up to 18 percentage points. The longer the queue of critical care patients waiting for ICU treatment and the larger the maximum number of patients subject to tertiary triage, the greater the effect on the reduction of mortality in the ICU.

CONCLUSION

A SAPS II score-based tertiary triage policy was superior in our simulation model. Random allocation or "first come, first served" policies yield the lowest survival rates, as will adherence to the new German Triage Act. An interdisciplinary discussion including an ethical and legal perspective is important for the social interpretation of our data-driven results.

Effectiveness of an Algorithmic Approach to Ventilator Withdrawal at the End of Life: A Stepped Wedge Cluster Randomized Trial

Background: The transition to spontaneous breathing puts patients who are undergoing ventilator withdrawal at high risk for developing respiratory distress. A patient-centered algorithmic approach could standardize this process and meet unique patient needs because a single approach (weaning vs. one-step extubation) does not capture the needs of a heterogenous population undergoing this palliative procedure. Objectives: (1) Demonstrate that the algorithmic approach can be effective to ensure greater patient respiratory comfort compared to usual care; (2) determine differences in opioid or benzodiazepine use; (3) predict factors associated with duration of survival. Design/Settings/Measures: A stepped-wedge cluster randomized design at five sites was used. Sites crossed over to the algorithm in random order after usual care data were obtained. Patient comfort was measured with the Respiratory Distress Observation Scale© (RDOS) at baseline, at ventilator off, and every 15-minutes for an hour. Parenteral morphine and lorazepam equivalents from the onset of the process until patient death were calculated. Results: Usual care data n = 120, algorithm data n = 48. Gender and race were evenly distributed. All patients in the usual care arm underwent a one-step ventilator cessation; 58% of patients in the algorithm arm were weaned over an average of 18 ± 27 minutes as prescribed in the algorithm. Patients had significantly less respiratory distress in the intervention arm (F = 10.41, p = 0.0013, effective size [es] = 0.49). More opioids (t = -2.30, p = 0.023) and benzodiazepines (t = -2.08, p = 0.040) were given in the control arm. Conclusions: The algorithm was effective in ensuring patient respiratory comfort. Surprisingly, more medication was given in the usual care arm; however, less may be needed when distress is objectively measured (RDOS), and treatment is initiated as soon as distress develops as in the algorithm. Clinical Trial Registration number: NCT03121391.

Diagnostic yield of viral multiplex PCR during acute exacerbation of COPD admitted to the intensive care unit: a pilot study

Acute exacerbation of chronic obstructive pulmonary disease (AECOPD) is one of the leading causes of admission to the intensive care unit, often triggered by a respiratory tract infection of bacterial or viral aetiology. Managing antibiotic therapy in this context remains a challenge. Respiratory panel molecular tests allow identifying viral aetiologies of AECOPD. We hypothesized that the systematic use of a respiratory multiplex PCR (mPCR) would help antibiotics saving in severe AECOPD. Our objectives were to describe the spectrum of infectious aetiologies of severe AECOPD, using a diagnostic approach combining conventional diagnostic tests and mPCR, and to measure antibiotics exposure. The study was bicentric, prospective, observational, and included 105 critically ill patients with a severe AECOPD of presumed infectious aetiology, in whom a respiratory mPCR with a viral panel was performed in addition to conventional microbiological tests. Altogether, the microbiological documentation rate was 50%, including bacteria alone (19%), respiratory viruses alone (16%), and mixed viruses and bacterial species (16%). The duration of antibiotic therapy was shorter in patients without documented bacterial infection (5.6 vs. 9 days; P = 0.0006). This pilot study suggests that molecular tests may help for the proper use of anti-infective treatments in critically ill patients with severe AECOPD.

Role of gut microbiota and bacterial translocation in acute intestinal injury and mortality in patients admitted in ICU for septic shock

Introduction

Sepsis is a life-threatening organ dysfunction with high mortality rate. The gut origin hypothesis of multiple organ dysfunction syndrome relates to loss of gut barrier function and the ensuing bacterial translocation. The aim of this study was to describe the evolution of gut microbiota in a cohort of septic shock patients over seven days and the potential link between gut microbiota and bacterial translocation.

Methods

Sixty consecutive adult patients hospitalized for septic shock in intensive care units (ICU) were prospectively enrolled. Non-inclusion criteria included patients with recent or scheduled digestive surgery, having taken laxatives, pre- or probiotic in the previous seven days, a progressive digestive neoplasia, digestive lymphoma, chronic inflammatory bowel disease, moribund patient, and pregnant and lactating patients. The primary objective was to evaluate the evolution of bacterial diversity and richness of gut microbiota during seven days in septic shock. Epidemiological, clinical and biological data were gathered over seven days. Gut microbiota was analyzed through a metagenomic approach. 100 healthy controls were selected among healthy blood donors for reference basal 16S rDNA values.

Results

Significantly lower bacterial diversity and richness was observed in gut microbiota of patients at Day 7 compared with Day 0 (p<0.01). SOFA score at Day 0, Acute Gastrointestinal Injury (AGI) local grade, septic shock origin and bacterial translocation had an impact on alpha diversity. A large increase in Enterococcus genus was observed at Day 7 with a decrease in Enterobacterales, Clostridiales, Bifidobacterium and other butyrate-producing bacteria.

Discussion

This study shows the importance of bacterial translocation during AGI in septic shock patients. This bacterial translocation decreases during hospitalization in ICUs in parallel to the decrease of microbiota diversity. This work highlights the role of gut microbiota and bacterial translocation during septic shock.

Development and Validation of Machine Learning-Based Models to Predict In-Hospital Mortality in Life-Threatening Ventricular Arrhythmias: Retrospective Cohort Study

BACKGROUND

Life-threatening ventricular arrhythmias (LTVAs) are main causes of sudden cardiac arrest and are highly associated with an increased risk of mortality. A prediction model that enables early identification of the high-risk individuals is still lacking.

OBJECTIVE

We aimed to build machine learning (ML)-based models to predict in-hospital mortality in patients with LTVA.

METHODS

A total of 3140 patients with LTVA were randomly divided into training (n=2512, 80%) and internal validation (n=628, 20%) sets. Moreover, data of 2851 patients from another database were collected as the external validation set. The primary output was the probability of in-hospital mortality. The discriminatory ability was evaluated by the area under the receiver operating characteristic curve (AUC). The prediction performances of 5 ML algorithms were compared with 2 conventional scoring systems, namely, the simplified acute physiology score (SAPS-II) and the logistic organ dysfunction system (LODS).

RESULTS

The prediction performance of the 5 ML algorithms significantly outperformed the traditional models in predicting in-hospital mortality. CatBoost showed the highest AUC of 90.5% (95% CI 87.5%-93.5%), followed by LightGBM with an AUC of 90.1% (95% CI 86.8%-93.4%). Conversely, the predictive values of SAPS-II and LODS were unsatisfactory, with AUCs of 78.0% (95% CI 71.7%-84.3%) and 74.9% (95% CI 67.2%-82.6%), respectively. The superiority of ML-based models was also shown in the external validation set.

CONCLUSIONS

ML-based models could improve the predictive values of in-hospital mortality prediction for patients with LTVA compared with traditional scoring systems.

Hemodynamic Effects of Awake Prone Positioning With COVID-19 Acute Respiratory Failure

INTRODUCTION

Awake prone positioning (PP) reduces need for intubation for patients with COVID-19 with acute respiratory failure. We investigated the hemodynamic effects of awake PP in non-ventilated subjects with COVID-19 acute respiratory failure.

METHODS

We conducted a single-center prospective cohort study. Adult hypoxemic subjects with COVID-19 not requiring invasive mechanical ventilation receiving at least one PP session were included. Hemodynamic assessment was done with transthoracic echocardiography before, during, and after a PP session.

RESULTS

Twenty-six subjects were included. We observed a significant and reversible increase in cardiac index (CI) during PP compared to supine position (SP): 3.0 ± 0.8 L/min/m2 in PP, 2.5 ± 0.6 L/min/m2 before PP (SP1), and 2.6 ± 0.5 L/min/m2 after PP (SP2, P < .001). A significant improvement in right ventricular (RV) systolic function was also evidenced during PP: The RV fractional area change was 36 ± 10% in SP1, 46 ± 10% during PP, and 35 ± 8% in SP2 (P < .001). There was no significant difference in PaO2 /FIO2 and breathing frequency.

CONCLUSION

CI and RV systolic function are improved by awake PP in non-ventilated subjects with COVID-19 with acute respiratory failure.

Critically ill patient mortality by age: long-term follow-up (CIMbA-LT)

BACKGROUND

The past years have witnessed dramatic changes in the population admitted to the intensive care unit (ICU). Older and sicker patients are now commonly treated in this setting due to the newly available sophisticated life support. However, the short- and long-term benefit of this strategy is scarcely studied.

METHODS

The Critically Ill patients' mortality by age: Long-Term follow-up (CIMbA-LT) was a multicentric, nationwide, retrospective, observational study addressing short- and long-term prognosis of patients admitted to Portuguese multipurpose ICUs, during 4 years, according to their age and disease severity. Patients were followed for two years after ICU admission. The standardized hospital mortality ratio (SMR) was calculated according to the Simplified Acute Physiology Score (SAPS) II and the follow-up risk, for patients discharged alive from the hospital, according to official demographic national data for age and gender. Survival curves were plotted according to age group.

RESULTS

We included 37.118 patients, including 15.8% over 80 years old. The mean SAPS II score was 42.8 ± 19.4. The ICU all-cause mortality was 16.1% and 76% of all patients survive until hospital discharge. The SAPS II score overestimated hospital mortality [SMR at hospital discharge 0.7; 95% confidence interval (CI) 0.63-0.76] but accurately predicted one-year all-cause mortality [1-year SMR 1.01; (95% CI 0.98-1.08)]. Survival curves showed a peak in mortality, during the first 30 days, followed by a much slower survival decline thereafter. Older patients had higher short- and long-term mortality and their hospital SMR was also slightly higher (0.76 vs. 0.69). Patients discharged alive from the hospital had a 1-year relative mortality risk of 6.3; [95% CI 5.8-6.7]. This increased risk was higher for younger patients [21.1; (95% CI 15.1-39.6) vs. 2.4; (95% CI 2.2-2.7) for older patients].

CONCLUSIONS

Critically ill patients' mortality peaked in the first 30 days after ICU admission. Older critically ill patients had higher all-cause mortality, including a higher hospital SMR. A long-term increased relative mortality risk was noted in patients discharged alive from the hospital, but this was more noticeable in younger patients.

Predictive outcomes of APACHE II and expanded SAPS II mortality scoring systems in coronary care unit

OBJECTIVE

We investigated the predictive values of the expanded Simplified Acute Physiology Score (SAPS) II and Acute Physiologic Score and Chronic Health Evaluation (APACHE) II score in predicting in-hospital mortality in coronary care unit (CCU) patients.

METHODS

In this study, expanded SAPS II and APACHE II scores were calculated in the CCU of a single-center tertiary hospital. Patients admitted to CCU with any cardivascular indication were included in the study. Both scores were calculated according to previously determined criteria. Calibration and discrimination abilities of the scores in predicting in-hospital mortality were tested with Hosmer-Lemeshow goodness-of-fit C chi-square and receiver operating characteristics (ROC) curve analyses.

RESULTS

A total of 871 patients were included in the analysis. The goodness-of-fit C chi-square test showed that both scores have a good performance in predicting survivors and nonsurvivors in CCU. Expanded SAPS II score has a sensitivity of 80% and a specificity of 91.8% with the cut-off value of 5.55, while APACHE II has a sensitivity of 75.9% and a specificity of 87.4% with the cut-off value of 16.5 in predicting mortality.

CONCLUSION

Expanded SAPS II and APACHE II scores have good ability to predict in-hospital mortality in CCU patients. Therefore, they can be used as a tool to predict short-term mortality in cardiovascular emergencies.

Predicting neurological outcome after cardiac arrest by combining computational parameters extracted from standard and deviant responses from auditory evoked potentials

Background

Despite multimodal assessment (clinical examination, biology, brain MRI, electroencephalography, somatosensory evoked potentials, mismatch negativity at auditory evoked potentials), coma prognostic evaluation remains challenging.

Methods

We present here a method to predict the return to consciousness and good neurological outcome based on classification of auditory evoked potentials obtained during an oddball paradigm. Data from event-related potentials (ERPs) were recorded noninvasively using four surface electroencephalography (EEG) electrodes in a cohort of 29 post-cardiac arrest comatose patients (between day 3 and day 6 following admission). We extracted retrospectively several EEG features (standard deviation and similarity for standard auditory stimulations and number of extrema and oscillations for deviant auditory stimulations) from the time responses in a window of few hundreds of milliseconds. The responses to the standard and the deviant auditory stimulations were thus considered independently. By combining these features, based on machine learning, we built a two-dimensional map to evaluate possible group clustering.

Results

Analysis in two-dimensions of the present data revealed two separated clusters of patients with good versus bad neurological outcome. When favoring the highest specificity of our mathematical algorithms (0.91), we found a sensitivity of 0.83 and an accuracy of 0.90, maintained when calculation was performed using data from only one central electrode. Using Gaussian, K-neighborhood and SVM classifiers, we could predict the neurological outcome of post-anoxic comatose patients, the validity of the method being tested by a cross-validation procedure. Moreover, the same results were obtained with one single electrode (Cz).

Conclusion

statistics of standard and deviant responses considered separately provide complementary and confirmatory predictions of the outcome of anoxic comatose patients, better assessed when combining these features on a two-dimensional statistical map. The benefit of this method compared to classical EEG and ERP predictors should be tested in a large prospective cohort. If validated, this method could provide an alternative tool to intensivists, to better evaluate neurological outcome and improve patient management, without neurophysiologist assistance.

Development and validation of a deep learning model to predict the survival of patients in ICU

BACKGROUND

Patients in the intensive care unit (ICU) are often in critical condition and have a high mortality rate. Accurately predicting the survival probability of ICU patients is beneficial to timely care and prioritizing medical resources to improve the overall patient population survival. Models developed by deep learning (DL) algorithms show good performance on many models. However, few DL algorithms have been validated in the dimension of survival time or compared with traditional algorithms.

METHODS

Variables from the Early Warning Score, Sequential Organ Failure Assessment Score, Simplified Acute Physiology Score II, Acute Physiology and Chronic Health Evaluation (APACHE) II, and APACHE IV models were selected for model development. The Cox regression, random survival forest (RSF), and DL methods were used to develop prediction models for the survival probability of ICU patients. The prediction performance was independently evaluated in the MIMIC-III Clinical Database (MIMIC-III), the eICU Collaborative Research Database (eICU), and Shanghai Pulmonary Hospital Database (SPH).

RESULTS

Forty variables were collected in total for model development. 83 943 participants from 3 databases were included in the study. The New-DL model accurately stratified patients into different survival probability groups with a C-index of >0.7 in the MIMIC-III, eICU, and SPH, performing better than the other models. The calibration curves of the models at 3 and 10 days indicated that the prediction performance was good. A user-friendly interface was developed to enable the model's convenience.

CONCLUSIONS

Compared with traditional algorithms, DL algorithms are more accurate in predicting the survival probability during ICU hospitalization. This novel model can provide reliable, individualized survival probability prediction.

Learning to predict in-hospital mortality risk in the intensive care unit with attention-based temporal convolution network

BACKGROUND

Dynamic prediction of patient mortality risk in the ICU with time series data is limited due to high dimensionality, uncertainty in sampling intervals, and other issues. A new deep learning method, temporal convolution network (TCN), makes it possible to deal with complex clinical time series data in ICU. We aimed to develop and validate it to predict mortality risk using time series data from MIMIC III dataset.

METHODS

A total of 21,139 records of ICU stays were analysed and 17 physiological variables from the MIMIC III dataset were used to predict mortality risk. Then we compared the model performance of the attention-based TCN with that of traditional artificial intelligence (AI) methods.

RESULTS

The area under receiver operating characteristic (AUCROC) and area under precision-recall curve (AUC-PR) of attention-based TCN for predicting the mortality risk 48 h after ICU admission were 0.837 (0.824 -0.850) and 0.454, respectively. The sensitivity and specificity of attention-based TCN were 67.1% and 82.6%, respectively, compared to the traditional AI method, which had a low sensitivity (< 50%).

CONCLUSIONS

The attention-based TCN model achieved better performance in the prediction of mortality risk with time series data than traditional AI methods and conventional score-based models. The attention-based TCN mortality risk model has the potential for helping decision-making for critical patients.

TRIAL REGISTRATION

Data used for the prediction of mortality risk were extracted from the freely accessible MIMIC III dataset. The project was approved by the Institutional Review Boards of Beth Israel Deaconess Medical Center (Boston, MA) and the Massachusetts Institute of Technology (Cambridge, MA). Requirement for individual patient consent was waived because the project did not impact clinical care and all protected health information was deidentified. The data were accessed via a data use agreement between PhysioNet, a National Institutes of Health-supported data repository (https://www.physionet.org/), and one of us (Yu-wen Chen, Certification Number: 28341490). All methods were carried out in accordance with the institutional guidelines and regulations.

Early Prediction of Sepsis Onset Using Neural Architecture Search Based on Genetic Algorithms

Sepsis is a life-threatening condition with a high mortality rate. Early prediction and treatment are the most effective strategies for increasing survival rates. This paper proposes a neural architecture search (NAS) model to predict the onset of sepsis with a low computational cost and high search performance by applying a genetic algorithm (GA). The proposed model shares the weights of all possible connection nodes internally within the neural network. Externally, the search cost is reduced through the weight-sharing effect between the genotypes of the GA. A predictive analysis was performed using the Medical Information Mart for Intensive Care III (MIMIC-III), a medical time-series dataset, with the primary objective of predicting sepsis onset 3 h before occurrence. In addition, experiments were conducted under various prediction times (0-12 h) for comparison. The proposed model exhibited an area under the receiver operating characteristic curve (AUROC) score of 0.94 (95% CI: 0.92-0.96) for 3 h, which is 0.31-0.26 higher than the scores obtained using the Sequential Organ Failure Assessment (SOFA), quick SOFA (qSOFA), and Simplified Acute Physiology Score (SAPS) II scoring systems. Furthermore, the proposed model exhibited a 12% improvement in the AUROC value over a simple model based on the long short-term memory neural network. Additionally, it is not only optimally searchable for sepsis onset prediction, but also outperforms conventional models that use similar predictive purposes and datasets. Notably, it is sufficiently robust to shape changes in the input data and has less structural dependence.

Management and outcomes in critically ill nonagenarian versus octogenarian patients

Background

Intensive care unit (ICU) patients age 90 years or older represent a growing subgroup and place a huge financial burden on health care resources despite the benefit being unclear. This leads to ethical problems. The present investigation assessed the differences in outcome between nonagenarian and octogenarian ICU patients.

Methods

We included 7900 acutely admitted older critically ill patients from two large, multinational studies. The primary outcome was 30-day-mortality, and the secondary outcome was ICU-mortality. Baseline characteristics consisted of frailty assessed by the Clinical Frailty Scale (CFS), ICU-management, and outcomes were compared between octogenarian (80–89.9 years) and nonagenarian (> 90 years) patients. We used multilevel logistic regression to evaluate differences between octogenarians and nonagenarians.

Results

The nonagenarians were 10% of the entire cohort. They experienced a higher percentage of frailty (58% vs 42%; p < 0.001), but lower SOFA scores at admission (6 + 5 vs. 7 + 6; p < 0.001). ICU-management strategies were different. Octogenarians required higher rates of organ support and nonagenarians received higher rates of life-sustaining treatment limitations (40% vs. 33%; p < 0.001). ICU mortality was comparable (27% vs. 27%; p = 0.973) but a higher 30-day-mortality (45% vs. 40%; p = 0.029) was seen in the nonagenarians. After multivariable adjustment nonagenarians had no significantly increased risk for 30-day-mortality (aOR 1.25 (95% CI 0.90–1.74; p = 0.19)).

Conclusion

After adjustment for confounders, nonagenarians demonstrated no higher 30-day mortality than octogenarian patients. In this study, being age 90 years or more is no particular risk factor for an adverse outcome. This should be considered– together with illness severity and pre-existing functional capacity - to effectively guide triage decisions.

Trial registration

NCT03134807 and NCT03370692.

Improving the In-Hospital Mortality Prediction of Diabetes ICU Patients Using a Process Mining/Deep Learning Architecture

Diabetes intensive care unit (ICU) patients are at increased risk of complications leading to in-hospital mortality. Assessing the likelihood of death is a challenging and time consuming task due to a large number of influencing factors. Healthcare providers are interested in the detection of ICU patients at higher risk, such that risk factors can possibly be mitigated. While such severity scoring methods exist, they are commonly based on a snapshot of the health conditions of a patient during the ICU stay and do not specifically consider a patient's prior medical history. In this paper, a process mining/deep learning architecture is proposed to improve established severity scoring methods by incorporating the medical history of diabetes patients. First, health records of past hospital encounters are converted to event logs suitable for process mining. The event logs are then used to discover a process model that describes the past hospital encounters of patients. An adaptation of Decay Replay Mining is proposed to combine medical and demographic information with established severity scores to predict the in hospital mortality of diabetes ICU patients. Significant performance improvements are demonstrated compared to established risk severity scoring methods and machine learning approaches using the Medical Information Mart for Intensive Care III dataset.

Prediction of Mortality in Surgical Intensive Care Unit Patients Using Machine Learning Algorithms

Objective: Predicting prognosis of in-hospital patients is critical. However, it is challenging to accurately predict the life and death of certain patients at certain period. To determine whether machine learning algorithms could predict in-hospital death of critically ill patients with considerable accuracy and identify factors contributing to the prediction power.

Materials and Methods: Using medical data of 1,384 patients admitted to the Surgical Intensive Care Unit (SICU) of our institution, we investigated whether machine learning algorithms could predict in-hospital death using demographic, laboratory, and other disease-related variables, and compared predictions using three different algorithmic methods. The outcome measurement was the incidence of unexpected postoperative mortality which was defined as mortality without pre-existing not-for-resuscitation order that occurred within 30 days of the surgery or within the same hospital stay as the surgery.

Results: Machine learning algorithms trained with 43 variables successfully classified dead and live patients with very high accuracy. Most notably, the decision tree showed the higher classification results (Area Under the Receiver Operating Curve, AUC = 0.96) than the neural network classifier (AUC = 0.80). Further analysis provided the insight that serum albumin concentration, total prenatal nutritional intake, and peak dose of dopamine drug played an important role in predicting the mortality of SICU patients.

Conclusion: Our results suggest that machine learning algorithms, especially the decision tree method, can provide information on structured and explainable decision flow and accurately predict hospital mortality in SICU hospitalized patients.

Effect of early treatment with polyvalent immunoglobulin on acute respiratory distress syndrome associated with SARS-CoV-2 infections (ICAR trial): study protocol for a randomized controlled trial

BACKGROUND

As of mid-June 2020, 7,500,000 people were infected with SARS-CoV-2 worldwide and 420,000 people died, mainly from coronavirus disease 2019 (COVID-19)-related acute respiratory distress syndrome (ARDS). COVID-19-related ARDS is subject to a mortality rate of 50% and prolonged period of mechanical ventilation, with no specific pharmacological treatment currently available (Infection au nouveau Coronavirus (SARS-CoV-2), COVID-19, France et Monde. https://www.santepubliquefrance.fr/dossiers/coronavirus-covid-19 ). Because of its immunomodulatory action, we propose to evaluate the efficacy and safety of intravenous immunoglobulin (IVIG) administration in patients developing COVID-19-related ARDS.

METHODS

The trial is a phase III double-blind, randomized, multicenter, parallel group, concurrent, controlled study in hospitalized participants with COVID-19 requiring mechanical ventilation using a sequential design. Participants in the treatment group will receive infusions of polyvalent immunoglobulin for 4 consecutive days, and the placebo group will receive an equivalent volume of sodium chloride 0.9% for the same duration. The primary outcome is the number of ventilator-free days up to the 28th day. Secondary objectives are to evaluate the effect of IVIG on (1) organ failure according to the Sequential Organ Failure Assessment (SOFA) score at 14 and 28 days, (2) lung injury score at 14 and 28 days, (3) the occurrence of grade 3 or 4 adverse events of IVIG, (4) length of intensive care unit (ICU) stay, (5) length of hospital stay, (6) functional outcomes at day 90 defined by the activities of daily living and instrumental activities of the daily living scales, and (7) 90-day survival. One hundred thirty-eight subjects will be randomized in a 1:1 ratio to IVIG or placebo groups (69 in each group), considering 90% power, alpha level 0.05 (two sides), and 0.67 effect size level.

DISCUSSION

The ICAR trial investigates the effect of IVIG in COVID-19-related ARDS. We expect an increase in the survival rate and a reduction in the duration of mechanical ventilation, which is associated with significant morbidity.

TRIAL REGISTRATION

EudraCT 2020-001570-30. ClinicalTrials.gov NCT04350580 . Registered on 17 April 2020.

Pediatric Simplified Acute Physiology Score II: Establishment of a New, Repeatable Pediatric Mortality Risk Assessment Score

Objectives: In critical care it is crucial to appropriately assess the risk of mortality for each patient. This is especially relevant in pediatrics, with its need for accurate and repeatable scoring. Aim of this study was to evaluate an age-adapted version of the expanded Simplified Acute Physiology Score II; (p-SAPS II), a repeatable, newly-designed scoring system compared to established scores (Pediatric Sequential Organ Failure Assessment Score/pSOFA, Pediatric Logistic Organ Dysfunction Score-2/PELOD-2 and Pediatric Index of Mortality 3/PIM3). Design: This retrospective cohort pilot study included data collected from patients admitted to the Pediatric Intensive Care Unit (PICU) at the Medical University of Vienna between July 2017 through December 2018. Patients: 231 admissions were included, comprising neonates (gestational age of ≥ 37 weeks) and patients up to 18 years of age with a PICU stay longer than 48 h. Main Outcomes: Mortality risk prediction and discrimination between survivors and non-survivors were the main outcomes of this study. The primary statistical methods for evaluating the performance of each score were the area under the receiver operating characteristic curve (AUROC) and goodness-of-fit test. Results: Highest AUROC curve was calculated for p-SAPS II (AUC = 0.86; 95% CI: 0.77-0.96; p < 0.001). This was significantly higher than the AUROCs of PELOD-2/pSOFA but not of PIM3. However, in a logistic regression model including p-SAPS II and PIM3 as covariates, p-SAPS II had a significant effect on the accuracy of prediction (p = 0.003). Nevertheless, according to the goodness-of-fit test for p-SAPS II and PIM3, p-SAPS II overestimated the number of deaths, whereas PIM3 showed acceptable estimations. Repeatability testing showed increasing AUROC values for p-SAPS II throughout the clinical stay (0.96 at day 28) but still no significant difference to PIM 3. The prediction accuracy, although improved over the days and even exceeded PIM 3. Conclusions: The newly-created p-SAPS II performed better than the established PIM3 in terms of discriminating between survivors and non-survivors. Furthermore, p-SAPS II can be assessed repeatably throughout a patient's PICU stay what improves mortality prediction. However, there is still a need to optimize calibration of the score to accurately predict mortality sooner throughout the clinical stay.

Number of intensivists per bed is associated with efficiency of Dutch intensive care units

PURPOSE

To measure efficiency in Intensive Care Units (ICUs) and to determine which organizational factors are associated with ICU efficiency, taking confounding factors into account.

MATERIALS AND METHODS

We used data of all consecutive admissions to Dutch ICUs between January 1, 2016 and January 1, 2019 and recorded ICU organizational factors. We calculated efficiency for each ICU by averaging the Standardized Mortality Ratio (SMR) and Standardized Resource Use (SRU) and examined the relationship between various organizational factors and ICU efficiency. We thereby compared the results of linear regression models before and after covariate adjustment using propensity scores.

RESULTS

We included 164,399 admissions from 83 ICUs. ICU efficiency ranged from 0.51-1.42 (average 0.99, 0.15 SD). The unadjusted model as well as the propensity score adjusted model showed a significant association between the ratio of employed intensivists per ICU bed and ICU efficiency. Other organizational factors had no statistically significant association with ICU efficiency after adjustment.

CONCLUSIONS

We found marked variability in efficiency in Dutch ICUs. After applying covariate adjustment using propensity scores, we identified one organizational factor, ratio intensivists per bed, having an association with ICU efficiency.

Early abolition of cough reflex predicts mortality in deeply sedated brain-injured patients

Background

Deep sedation may hamper the detection of neurological deterioration in brain-injured patients. Impaired brainstem reflexes within the first 24 h of deep sedation are associated with increased mortality in non-brain-injured patients. Our objective was to confirm this association in brain-injured patients.

Methods

This was an observational prospective multicenter cohort study involving four neuro-intensive care units. We included acute brain-injured patients requiring deep sedation, defined by a Richmond Assessment Sedation Scale (RASS) < −3. Neurological assessment was performed at day 1 and included pupillary diameter, pupillary light, corneal and cough reflexes, and grimace and motor response to noxious stimuli. Pre-sedation Glasgow Coma Scale (GCS) and Simplified Acute Physiology Score (SAPS-II) were collected, as well as the cause of death in the Intensive Care Unit (ICU).

Results

A total of 137 brain-injured patients were recruited, including 70 (51%) traumatic brain-injured patients, 40 (29%) vascular (subarachnoid hemorrhage or intracerebral hemorrhage). Thirty patients (22%) died in the ICU. At day 1, the corneal (OR 2.69, p = 0.034) and cough reflexes (OR 5.12, p = 0.0003) were more frequently abolished in patients that died in the ICU. In a multivariate analysis, abolished cough reflex was associated with ICU mortality after adjustment to pre-sedation GCS, SAPS-II, RASS (OR: 5.19, 95% CI [1.92–14.1], p = 0.001) or dose of sedatives (OR: 8.89, 95% CI [2.64–30.0], p = 0.0004).

Conclusion

Early (day 1) cough reflex abolition is an independent predictor of mortality in deeply sedated brain-injured patients. Abolished cough reflex likely reflects a brainstem dysfunction that might result from the combination of primary and secondary neuro-inflammatory cerebral insults revealed and/or worsened by sedation.

Overall and subgroup specific performance of the pediatric index of mortality 2 score in Switzerland: a national multicenter study

Pediatric Index of Mortality (PIM) 2 score is used in pediatric intensive care unit (PICU) to predict the patients' risk of death. The performance of this model has never been assessed in Switzerland. The aim of this study was to evaluate the performance of the PIM2 score in the whole cohort and in pre-specified diagnostic subgroups of patients admitted to PICUs in Switzerland. All children younger than 16 years admitted to any PICU in Switzerland between January 1, 2012 and December 31, 2017 were included in the study. A total of 22,382 patients were analyzed. Observed mortality was 2%, whereas mortality predicted by PIM2 was 4.2% (SMR = 0.47, 95% CI, 0.42-0.52). Calibration was also poor across the deciles of mortality risks (p < 0.001). The AUC-ROC for the entire cohort was 0.88 (95% CI, 0.87-0.90). Calibration varied significantly according to primary diagnosis.Conclusion: The performance of the PIM 2 score in a cohort of Swiss patients is poor with adequate discrimination and poor calibration. The PIM 2 score tends to under predict the number of deaths among septic patients and in patients admitted after a cardiorespiratory arrest. What is Known: •PIM2 score is a widely used mortality prediction model in PICU. •PIM2 performance among uncommon but clinically relevant diagnostic subgroups of patients is unknown. •The performance of PIM2 score has never been assessed in Switzerland. What is New: •The performance of the PIM 2 score in a cohort of Swiss patients is poor with adequate discrimination and poor calibration. •Calibration varies significantly according to primary diagnosis. The PIM 2 score under predict the number of deaths among septic patients and in patients admitted after a cardiorespiratory arrest.

Adjusting for Disease Severity Across ICUs in Multicenter Studies

Supplemental Digital Content is available in the text.

Objectives:

To compare methods to adjust for confounding by disease severity during multicenter intervention studies in ICU, when different disease severity measures are collected across centers.

Design:

In silico simulation study using national registry data.

Setting:

Twenty mixed ICUs in The Netherlands.

Subjects:

Fifty-five–thousand six-hundred fifty-five ICU admissions between January 1, 2011, and January 1, 2016.

Interventions:

None.

Measurements and Main Results:

To mimic an intervention study with confounding, a fictitious treatment variable was simulated whose effect on the outcome was confounded by Acute Physiology and Chronic Health Evaluation IV predicted mortality (a common measure for disease severity). Diverse, realistic scenarios were investigated where the availability of disease severity measures (i.e., Acute Physiology and Chronic Health Evaluation IV, Acute Physiology and Chronic Health Evaluation II, and Simplified Acute Physiology Score II scores) varied across centers. For each scenario, eight different methods to adjust for confounding were used to obtain an estimate of the (fictitious) treatment effect. These were compared in terms of relative (%) and absolute (odds ratio) bias to a reference scenario where the treatment effect was estimated following correction for the Acute Physiology and Chronic Health Evaluation IV scores from all centers. Complete neglect of differences in disease severity measures across centers resulted in bias ranging from 10.2% to 173.6% across scenarios, and no commonly used methodology—such as two-stage modeling or score standardization—was able to effectively eliminate bias. In scenarios where some of the included centers had (only) Acute Physiology and Chronic Health Evaluation II or Simplified Acute Physiology Score II available (and not Acute Physiology and Chronic Health Evaluation IV), either restriction of the analysis to Acute Physiology and Chronic Health Evaluation IV centers alone or multiple imputation of Acute Physiology and Chronic Health Evaluation IV scores resulted in the least amount of relative bias (0.0% and 5.1% for Acute Physiology and Chronic Health Evaluation II, respectively, and 0.0% and 4.6% for Simplified Acute Physiology Score II, respectively). In scenarios where some centers used Acute Physiology and Chronic Health Evaluation II, regression calibration yielded low relative bias too (relative bias, 12.4%); this was not true if these same centers only had Simplified Acute Physiology Score II available (relative bias, 54.8%).

Conclusions:

When different disease severity measures are available across centers, the performance of various methods to control for confounding by disease severity may show important differences. When planning multicenter studies, researchers should make contingency plans to limit the use of or properly incorporate different disease measures across centers in the statistical analysis.

Contribution of Capillary Refilling Time and Skin Mottling Score to Predict ICU Admission of Patients with Septic or haemorrhagic Shock Admitted to the Emergency Department: A TRCMARBSAU Study

Objective

In the emergency department (ED), the severity assessment of shock is a fundamental step prior to the admission in the intensive care unit (ICU). As biomarkers are time consuming to evaluate the severity of micro- and macro-circulation alteration, capillary refill time and skin mottling score are two simple, available clinical criteria validated to predict mortality in the ICU. The aim of the present study is to provide clinical evidence that capillary refill time and skin mottling score assessed in the ED also predict ICU admission of patients with septic or haemorrhagic shock.

Methods

This trial is an observational, non-randomised controlled study. A total of 1500 patients admitted to the ED for septic or haemorrhagic shock will be enrolled into the study. The primary outcome is the admission to the ICU.

Results

The study will not impact the treatments provided to each patient. Capillary refill time and skin mottling score will not be taken into account to decide patient's treatments and/or ICU admission. Patients will be followed up during their hospital stay to determine their precise destination after the ED (home, ICU or ward) and the 28- and 90-day mortality after hospital admission.

Conclusion

The results from the present study will provide clinical evidence on the correlation between the ICU admission and the capillary refill time and the skin mottling score in septic or haemorrhagic shock admitted to the ED. The aim of the present study is to provide two simple, reliable and non-invasive tools for the triage and early orientation of these patients.

Ethical considerations about artificial intelligence for prognostication in intensive care

Background

Prognosticating the course of diseases to inform decision-making is a key component of intensive care medicine. For several applications in medicine, new methods from the field of artificial intelligence (AI) and machine learning have already outperformed conventional prediction models. Due to their technical characteristics, these methods will present new ethical challenges to the intensivist.

Results

In addition to the standards of data stewardship in medicine, the selection of datasets and algorithms to create AI prognostication models must involve extensive scrutiny to avoid biases and, consequently, injustice against individuals or groups of patients. Assessment of these models for compliance with the ethical principles of beneficence and non-maleficence should also include quantification of predictive uncertainty. Respect for patients’ autonomy during decision-making requires transparency of the data processing by AI models to explain the predictions derived from these models. Moreover, a system of continuous oversight can help to maintain public trust in this technology. Based on these considerations as well as recent guidelines, we propose a pathway to an ethical implementation of AI-based prognostication. It includes a checklist for new AI models that deals with medical and technical topics as well as patient- and system-centered issues.

Conclusion

AI models for prognostication will become valuable tools in intensive care. However, they require technical refinement and a careful implementation according to the standards of medical ethics.

Contribution of the Pre-Hospital Blood Lactate Level in the Pre-Hospital Orientation of Septic Shock: The LAPHSUS Study

Objective

In the pre-hospital setting, the assessment of septic shock severity is essential when determining the optimal initial in-hospital level of care. As clinical signs can be faulted, there is a need for an additional component to enhance the severity assessment and to decide on in-hospital admission in the intensive care unit (ICU) or in the emergency department (ED). Point-of-care medical devices by yielding blood lactate value since the pre-hospital setting may give an easy and valuable component for the severity assessment and decision-making. The aim of this study is to provide clinical evidence that the pre-hospital blood lactate level predicts the 30-day mortality in patients with septic shock.

Methods

This trial is a prospective, observational, non-randomised controlled study. A total of 1,000 patients requiring a mobile ICU intervention for septic shock in the pre-hospital setting will be included. Pre-hospital blood lactate levels will not be taken into account to decide patients’ treatments and/or ED or ICU admission. In the pre-hospital setting, each patient will benefit from two measurements of the blood lactate level: initial measurement at the first contact, and final measurement at the hospital admission with a specific point-of-care medical device.

Conclusion

This study could provide clinical evidence that the pre-hospital blood lactate level predicts the 30-day mortality of patients with septic shock. The results from this study could also prove the utility of the pre-hospital blood lactate level for the triage and early orientation of patients with septic shock.

Sex differences in crude mortality rates and predictive value of intensive care unit-based scores when applied to the cardiac intensive care unit

BACKGROUND

Limited data exists regarding sex differences in outcome and predictive accuracy of intensive care unit-based scoring systems when applied to cardiac intensive care unit patients.

METHODS

We reviewed medical records of patients admitted to cardiac intensive care unit from 1 January 2011-31 December 2016. Sex differences in mortality rates and the performance of intensive care unit-based scoring systems in predicting in-hospital mortality were analyzed. Calibration was assessed by the Hosmer-Lemeshow test and locally weighted scatterplot smoothing curves. Discrimination was assessed using the c statistic and receiver-operating characteristic curve.

RESULTS

Among 6963 patients, 2713 (39%) were women. Overall in-hospital and cardiac intensive care unit mortality rates were similar in women and men (9.1% vs 9.4%, p=0.67 and 5.9% vs 6%, p=0.88, respectively) and in age and major diagnosis subgroups. Of the scoring systems, Acute Physiology and Chronic Health Evaluation III and Sequential Organ Failure Assessment had poor calibration (Hosmer-Lemeshow p value <0.001), while Simplified Acute Physiology Score II performed better (Hosmer-Lemeshow p value 0.09), in both women and men. All scores had good discrimination (C statistics >0.8). In the subgroups of acute myocardial infarction and heart failure patients, all scores had good calibration (Hosmer-Lemeshow p>0.001) and discrimination (C statistic >0.8) while in diagnosis subgroups with highest mortality, the calibration varied among scores and by sex, and discrimination was poor.

CONCLUSIONS

No sex differences in mortality were seen in cardiac intensive care unit patients. The mortality predictive value of intensive care unit-based scores is limited in both sexes and variable among different subgroups of diagnoses.

Predicting hospital mortality for intensive care unit patients: Time-series analysis

Current mortality prediction models and scoring systems for intensive care unit patients are generally usable only after at least 24 or 48 h of admission, as some parameters are unclear at admission. However, some of the most relevant measurements are available shortly following admission. It is hypothesized that outcome prediction may be made using information available in the earliest phase of intensive care unit admission. This study aims to investigate how early hospital mortality can be predicted for intensive care unit patients. We conducted a thorough time-series analysis on the performance of different data mining methods during the first 48 h of intensive care unit admission. The results showed that the discrimination power of the machine-learning classification methods after 6 h of admission outperformed the main scoring systems used in intensive care medicine (Acute Physiology and Chronic Health Evaluation, Simplified Acute Physiology Score and Sequential Organ Failure Assessment) after 48 h of admission.

Predicting in-hospital mortality of patients with acute kidney injury in the ICU using random forest model

OBJECTIVES

We aimed to construct a mortality prediction model using the random forest (RF) algorithm for acute kidney injury (AKI) patients in the intensive care unit (ICU), and compared its performance with that of two other machine learning models and the customized simplified acute physiology score (SAPS) II model.

METHODS

We used medical information mart for intensive care (MIMIC) III database for model development and performance comparison. The RF model uses the same predictor variable set as that of the SAPS II model. We also developed three other models and compared the RF model with the other three models in prediction performance. Three other models include support vector machine (SVM) model, artificial neural network (ANN) model and customized SAPS II model. In model comparison, the prediction performance of each model was assessed by the Brier score, the area under the receiver operating characteristic curve (AUROC), accuracy and F₁ score.

RESULTS

The final cohort consisted of 19044 patients with AKI in the ICU. The observed in-hospital mortality of AKI patients is 13.6% in the final cohort. The results of model performance comparison show that the Brier score of the RF model is 0.085 (95%CI: 0.084-0.086) and AUROC of the RF model is 0.866 (95%CI: 0.862-0.870). The accuracy of the RF model is 0.728 (95%CI: 0.715-0.741). The F₁ score of the RF model is 0.459 (95%CI: 0.449-0.470). The calibration plots show that the RF model slightly overestimates mortality in patients with low risk of death and underestimates mortality in patients with high risk of death.

CONCLUSION

There is great potential for the RF model in mortality prediction for AKI patients in ICU. The RF model may be helpful to aid ICU clinicians to make timely clinical intervention decisions for AKI patients, which is critical to help reduce the in-hospital mortality of AKI patients. A prospective study is necessary to evaluate the clinical utility of the RF model.

Clinical implications of the initial SAPS II in veno-arterial extracorporeal oxygenation

Background

Prediction of survival and weaning probability in VA ECMO (veno-arterial extracorporeal membrane oxygenation) patients could be of great benefit for real-time decision making on VA ECMO initiation in critical ill patients. We investigated whether the SAPS II score would be a real-time determinant for VA ECMO initiation and could be a predictor of survival and weaning probability in patients on VA ECMO.

Methods

Between January 1, 2010 and December 31, 2014, VA ECMO was carried out on 135 adult patients suffering from primary cardiogenic shock. To avoid selection bias, we excluded respiratory failure patients treated with VV or other types of ECMO. Successful VA ECMO weaning was defined as weaning, followed by stable survival for more than 48 hours. Survival after VA ECMO was defined as successful weaning and treatment of the underlying medical condition, followed by discharge without any further events.

Results

A total of 135 patients consisted of 41 women and 94 men, with a mean age of 59.4±16.5 years. Fifty-three patients had successful weaning, and 35 survived and were discharged uneventfully. Compared to the non-survivors, the survivors showed a lower SAPS II (67.77±20.79 vs. 90.29±13.31, P<0.001), a lower SOFA score (12.63±3.49 vs. 15.33±2.28, P<0.001), a lower predicted death rate (71.12±30.51 vs. 94.00±9.36, P<0.001), a higher initial ipH (7.14±0.22 vs. 6.98±0.15, P<0.001), and a lower initial lactate level (7.09±4.93 vs. 12.11±4.84, P<0.001). The average duration of hospital stay in the successful vs. failed weaning groups was 33.43±27.41 vs. 6.35±8.71 days, and the average duration of ICU stay in the successful vs. failed weaning groups was 20.60±16.88 vs. 5.39±5.95 days. By multivariate logistic regression analysis of initial parameters for VA ECMO assistance, the simplified acute physiology score II (SAPS II) (OR =1.1019, P=0.0389), ipH (OR =0.0010, P=0.0452), and hospital stay (OR =0.8140, P=0.001) had an association with in-hospital mortality on VA ECMO. The initial SAPS II score [area under the curve (AUC) =0.821] demonstrated significantly superior prediction of VA ECMO mortality than age (AUC =0.697), SOFA score (AUC =0.701), ipH (AUC =0.551), and the other parameters. By multivariable CoX regression analysis of survival, only the SAPS II score proved to have statistical significance (hazard ratio, 1.0423; 95% CI, 1.0083-1.0775; P=0.01).

Conclusions

Although the precise predictive scoring systems for VA ECMO still remains one of the most difficult challenges to ECMO physicians, the SAPS II score could provide valuable information on prognosis to patient himself, family members and caretakers, and might help physicians increase the survival rate and might avoid a waste of healthcare resources.

Early Hospital Mortality Prediction using Vital Signals

Early hospital mortality prediction is critical as intensivists strive to make efficient medical decisions about the severely ill patients staying in intensive care units (ICUs). As a result, various methods have been developed to address this problem based on clinical records. However, some of the laboratory test results are time-consuming and need to be processed. In this paper, we propose a novel method to predict mortality using features extracted from the heart signals of patients within the first hour of ICU admission. In order to predict the risk, quantitative features have been computed based on the heart rate signals of ICU patients suffering cardiovascular diseases. Each signal is described in terms of 12 statistical and signal-based features. The extracted features are fed into eight classifiers: decision tree, linear discriminant, logistic regression, support vector machine (SVM), random forest, boosted trees, Gaussian SVM, and K-nearest neighborhood (K-NN). To derive insight into the performance of the proposed method, several experiments have been conducted using the well-known clinical dataset named Medical Information Mart for Intensive Care III (MIMIC-III). The experimental results demonstrate the capability of the proposed method in terms of precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC). The decision tree classifier satisfies both accuracy and interpretability better than the other classifiers, producing an F1-score and AUC equal to 0.91 and 0.93, respectively. It indicates that heart rate signals can be used for predicting mortality in patients in the care units especially coronary care units (CCUs), achieving a comparable performance with existing predictions that rely on high dimensional features from clinical records which need to be processed and may contain missing information.

Decontamination Strategies and Bloodstream Infections With Antibiotic-Resistant Microorganisms in Ventilated Patients: A Randomized Clinical Trial

Importance

The effects of chlorhexidine (CHX) mouthwash, selective oropharyngeal decontamination (SOD), and selective digestive tract decontamination (SDD) on patient outcomes in ICUs with moderate to high levels of antibiotic resistance are unknown.

Objective

To determine associations between CHX 2%, SOD, and SDD and the occurrence of ICU-acquired bloodstream infections with multidrug-resistant gram-negative bacteria (MDRGNB) and 28-day mortality in ICUs with moderate to high levels of antibiotic resistance.

Design, Setting, and Participants

Randomized trial conducted from December 1, 2013, to May 31, 2017, in 13 European ICUs where at least 5% of bloodstream infections are caused by extended-spectrum β-lactamase-producing Enterobacteriaceae. Patients with anticipated mechanical ventilation of more than 24 hours were eligible. The final date of follow-up was September 20, 2017.

Interventions

Standard care was daily CHX 2% body washings and a hand hygiene improvement program. Following a baseline period from 6 to 14 months, each ICU was assigned in random order to 3 separate 6-month intervention periods with either CHX 2% mouthwash, SOD (mouthpaste with colistin, tobramycin, and nystatin), or SDD (the same mouthpaste and gastrointestinal suspension with the same antibiotics), all applied 4 times daily.

Main Outcomes and Measures

The occurrence of ICU-acquired bloodstream infection with MDRGNB (primary outcome) and 28-day mortality (secondary outcome) during each intervention period compared with the baseline period.

Results

A total of 8665 patients (median age, 64.1 years; 5561 men [64.2%]) were included in the study (2251, 2108, 2224, and 2082 in the baseline, CHX, SOD, and SDD periods, respectively). ICU-acquired bloodstream infection with MDRGNB occurred among 144 patients (154 episodes) in 2.1%, 1.8%, 1.5%, and 1.2% of included patients during the baseline, CHX, SOD, and SDD periods, respectively. Absolute risk reductions were 0.3% (95% CI, -0.6% to 1.1%), 0.6% (95% CI, -0.2% to 1.4%), and 0.8% (95% CI, 0.1% to 1.6%) for CHX, SOD, and SDD, respectively, compared with baseline. Adjusted hazard ratios were 1.13 (95% CI, 0.68-1.88), 0.89 (95% CI, 0.55-1.45), and 0.70 (95% CI, 0.43-1.14) during the CHX, SOD, and SDD periods, respectively, vs baseline. Crude mortality risks on day 28 were 31.9%, 32.9%, 32.4%, and 34.1% during the baseline, CHX, SOD, and SDD periods, respectively. Adjusted odds ratios for 28-day mortality were 1.07 (95% CI, 0.86-1.32), 1.05 (95% CI, 0.85-1.29), and 1.03 (95% CI, 0.80-1.32) for CHX, SOD, and SDD, respectively, vs baseline.

Conclusions and Relevance

Among patients receiving mechanical ventilation in ICUs with moderate to high antibiotic resistance prevalence, use of CHX mouthwash, SOD, or SDD was not associated with reductions in ICU-acquired bloodstream infections caused by MDRGNB compared with standard care.

Trial Registration

ClinicalTrials.gov Identifier: NCT02208154.

Decontamination Strategies and Bloodstream Infections With Antibiotic-Resistant Microorganisms in Ventilated Patients: A Randomized Clinical Trial

IMPORTANCE

The effects of chlorhexidine (CHX) mouthwash, selective oropharyngeal decontamination (SOD), and selective digestive tract decontamination (SDD) on patient outcomes in ICUs with moderate to high levels of antibiotic resistance are unknown.

OBJECTIVE

To determine associations between CHX 2%, SOD, and SDD and the occurrence of ICU-acquired bloodstream infections with multidrug-resistant gram-negative bacteria (MDRGNB) and 28-day mortality in ICUs with moderate to high levels of antibiotic resistance.

DESIGN, SETTING, AND PARTICIPANTS

Randomized trial conducted from December 1, 2013, to May 31, 2017, in 13 European ICUs where at least 5% of bloodstream infections are caused by extended-spectrum β-lactamase-producing Enterobacteriaceae. Patients with anticipated mechanical ventilation of more than 24 hours were eligible. The final date of follow-up was September 20, 2017.

INTERVENTIONS

Standard care was daily CHX 2% body washings and a hand hygiene improvement program. Following a baseline period from 6 to 14 months, each ICU was assigned in random order to 3 separate 6-month intervention periods with either CHX 2% mouthwash, SOD (mouthpaste with colistin, tobramycin, and nystatin), or SDD (the same mouthpaste and gastrointestinal suspension with the same antibiotics), all applied 4 times daily.

MAIN OUTCOMES AND MEASURES

The occurrence of ICU-acquired bloodstream infection with MDRGNB (primary outcome) and 28-day mortality (secondary outcome) during each intervention period compared with the baseline period.

RESULTS

A total of 8665 patients (median age, 64.1 years; 5561 men [64.2%]) were included in the study (2251, 2108, 2224, and 2082 in the baseline, CHX, SOD, and SDD periods, respectively). ICU-acquired bloodstream infection with MDRGNB occurred among 144 patients (154 episodes) in 2.1%, 1.8%, 1.5%, and 1.2% of included patients during the baseline, CHX, SOD, and SDD periods, respectively. Absolute risk reductions were 0.3% (95% CI, -0.6% to 1.1%), 0.6% (95% CI, -0.2% to 1.4%), and 0.8% (95% CI, 0.1% to 1.6%) for CHX, SOD, and SDD, respectively, compared with baseline. Adjusted hazard ratios were 1.13 (95% CI, 0.68-1.88), 0.89 (95% CI, 0.55-1.45), and 0.70 (95% CI, 0.43-1.14) during the CHX, SOD, and SDD periods, respectively, vs baseline. Crude mortality risks on day 28 were 31.9%, 32.9%, 32.4%, and 34.1% during the baseline, CHX, SOD, and SDD periods, respectively. Adjusted odds ratios for 28-day mortality were 1.07 (95% CI, 0.86-1.32), 1.05 (95% CI, 0.85-1.29), and 1.03 (95% CI, 0.80-1.32) for CHX, SOD, and SDD, respectively, vs baseline.

CONCLUSIONS AND RELEVANCE

Among patients receiving mechanical ventilation in ICUs with moderate to high antibiotic resistance prevalence, use of CHX mouthwash, SOD, or SDD was not associated with reductions in ICU-acquired bloodstream infections caused by MDRGNB compared with standard care.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02208154.

Machine learning in critical care: state-of-the-art and a sepsis case study

BACKGROUND

Like other scientific fields, such as cosmology, high-energy physics, or even the life sciences, medicine and healthcare face the challenge of an extremely quick transformation into data-driven sciences. This challenge entails the daunting task of extracting usable knowledge from these data using algorithmic methods. In the medical context this may for instance realized through the design of medical decision support systems for diagnosis, prognosis and patient management. The intensive care unit (ICU), and by extension the whole area of critical care, is becoming one of the most data-driven clinical environments.

RESULTS

The increasing availability of complex and heterogeneous data at the point of patient attention in critical care environments makes the development of fresh approaches to data analysis almost compulsory. Computational Intelligence (CI) and Machine Learning (ML) methods can provide such approaches and have already shown their usefulness in addressing problems in this context. The current study has a dual goal: it is first a review of the state-of-the-art on the use and application of such methods in the field of critical care. Such review is presented from the viewpoint of the different subfields of critical care, but also from the viewpoint of the different available ML and CI techniques. The second goal is presenting a collection of results that illustrate the breath of possibilities opened by ML and CI methods using a single problem, the investigation of septic shock at the ICU.

CONCLUSION

We have presented a structured state-of-the-art that illustrates the broad-ranging ways in which ML and CI methods can make a difference in problems affecting the manifold areas of critical care. The potential of ML and CI has been illustrated in detail through an example concerning the sepsis pathology. The new definitions of sepsis and the relevance of using the systemic inflammatory response syndrome (SIRS) in its diagnosis have been considered. Conditional independence models have been used to address this problem, showing that SIRS depends on both organ dysfunction measured through the Sequential Organ Failure (SOFA) score and the ICU outcome, thus concluding that SIRS should still be considered in the study of the pathophysiology of Sepsis. Current assessment of the risk of dead at the ICU lacks specificity. ML and CI techniques are shown to improve the assessment using both indicators already in place and other clinical variables that are routinely measured. Kernel methods in particular are shown to provide the best performance balance while being amenable to representation through graphical models, which increases their interpretability and, with it, their likelihood to be accepted in medical practice.

Driving pressure and acute respiratory distress syndrome in critically ill patients

BACKGROUND AND OBJECTIVE

Elevated driving pressure (ΔP) may be associated with increased risk of acute respiratory distress syndrome (ARDS) in patients admitted via the emergency department and with post-operative pulmonary complications in surgical patients. This study investigated the association of higher ΔP with the onset of ARDS in a high-risk, intensive care unit (ICU) population.

METHODS

This is a secondary analysis of a prospective multicentre observational study. Data for this ancillary study were obtained from intubated adult patients with at least one ARDS risk factor upon ICU admission enrolled in a previous multicentre observational study. Patients were followed up for the development of ARDS within 7 days (primary outcome). Univariate and multivariate analyses tested the association between ΔP (measured at ICU admission (baseline) or 24 h later (day 1)) and the development of ARDS.

RESULTS

A total of 221 patients were included in this study, among whom 34 (15%) developed ARDS within 7 days. These patients had higher baseline ΔP than those who did not (mean ± SD: 12.5 ± 3.1 vs 9.8 ± 3.4 cm H₂ O, respectively, P = 0.0001). The association between baseline ΔP and the risk of developing ARDS was robust to adjustment for baseline tidal volume, positive-end expiratory pressure, illness severity, serum lactate and sepsis, pneumonia, severe trauma and shock as primary ARDS risk factors (odds ratio: 1.20; 95% CI: 1.03-1.41; P = 0.02). The same results were found with day 1 ΔP.

CONCLUSION

Among at-risk ICU patients, higher ΔP may identify those who are more likely to develop ARDS.

Mismatch negativity to predict subsequent awakening in deeply sedated critically ill patients

BACKGROUND

Mismatch negativity (MMN) is the neurophysiological correlate of cognitive integration of novel stimuli. Although MMN is a well-established predictor of awakening in non-sedated comatose patients, its prognostic value in deeply sedated critically ill patients remains unknown. The aim of this prospective, observational pilot study was to investigate the prognostic value of MMN for subsequent awakening in deeply sedated critically ill patients.

METHODS

MMN was recorded in 43 deeply sedated critically ill patients on Day 3 of ICU admission using a classical 'odd-ball' paradigm that delivers rare deviant sounds in a train of frequent standard sounds. Individual visual analyses and a group level analysis of recordings were performed. MMN amplitudes were then analysed according to the neurological status (awake vs not awake) at Day 28.

RESULTS

Median (inter-quartile range) Richmond Assessment Sedation Scale (RASS) at the time of recording was -5 (range, from -5 to -4.5). Visual detection of MMN revealed a poor inter-rater agreement [kappa=0.17, 95% confidence interval (0.07-0.26)]. On Day 28, 30 (70%) patients had regained consciousness while 13 (30%) had not. Quantitative group level analysis revealed a significantly greater MMN amplitude for patients who awakened compared with those who had not [mean (standard deviation) = -0.65 (1.4) vs 0.08 (0.17) μV, respectively; P=0.003).

CONCLUSIONS

MMN can be observed in deeply sedated critically ill patients and could help predict subsequent awakening. However, visual analysis alone is unreliable and should be systematically completed with individual level statistics.

Clinical and Biological Predictors of Plasma Levels of Soluble RAGE in Critically Ill Patients: Secondary Analysis of a Prospective Multicenter Observational Study

Rationale

Although soluble forms of the receptor for advanced glycation end products (RAGE) have been recently proposed as biomarkers in multiple acute or chronic diseases, few studies evaluated the influence of usual clinical and biological parameters, or of patient characteristics and comorbidities, on circulating levels of soluble RAGE in the intensive care unit (ICU) setting.

Objectives

To determine, among clinical and biological parameters that are usually recorded upon ICU admission, which variables, if any, could be associated with plasma levels of soluble RAGE.

Methods

Data for this ancillary study were prospectively obtained from adult patients with at least one ARDS risk factor upon ICU admission enrolled in a large multicenter observational study. At ICU admission, plasma levels of total soluble RAGE (sRAGE) and endogenous secretory (es)RAGE were measured by duplicate ELISA and baseline patient characteristics, comorbidities, and usual clinical and biological indices were recorded. After univariate analyses, significant variables were used in multivariate, multidimensional analyses.

Measurements and Main Results

294 patients were included in this ancillary study, among whom 62% were admitted for medical reasons, including septic shock (11%), coma (11%), and pneumonia (6%). Although some variables were associated with plasma levels of RAGE soluble forms in univariate analysis, multidimensional analyses showed no significant association between admission parameters and baseline plasma sRAGE or esRAGE.

Conclusions

We found no obvious association between circulating levels of soluble RAGE and clinical and biological indices that are usually recorded upon ICU admission. This trial is registered with NCT02070536.

Changes in biopsychosocial outcomes for a mixed cohort of ICU survivors

Background

Prolonged inflammation and infection associated with being critically ill and the ensuing physical inactivity has proven negative effects on the recovery of physical function, psychological health and reintegration into society for intensive care unit (ICU) survivors. Limited evidence is available on changes in biopsychosocial outcomes for South Africans recovering from an episode of critical illness.

Objectives

To determine changes in biopsychosocial outcomes for a mixed cohort of ICU survivors in hospital and at 1 month and 6 months after discharge.

Method

A prospective, observational, longitudinal study was conducted. Severity of illness, mechanical ventilation (MV) duration and ICU and hospital length of stay (LOS) were recorded. Physical function in ICU test-scored (PFIT-s) was performed at discharge from ICU and hospital. At 1 month and 6 months, peripheral muscle strength, exercise endurance, health-related quality of life (HRQOL), depression status and return to work were assessed. Descriptive and inferential statistics were used.

Results

Participants (n = 24) had a median age of 51.5 years, majority were male (n = 19; 79%) and most were employed before admission (n = 20; 83%). At 6 months, 11 participants (n = 11) were part of the final sample. Median PFIT-s changed significantly (0.3 points; p = 0.02) between ICU and hospital discharge. Peripheral muscle strength improved significantly for upper and lower limbs over 6 months (p = 0.00–0.03) but change in median 6-minute walk test distance (65m) was not significantly different. Significant improvements occurred in mean Medical Outcomes Short Form-36 (SF-36) physical health component scores (8.8 ± 7.6; p = 0.00). Mean SF-36 mental health component scores had a strong negative relationship with MV duration (r = −0.7; p = 0.01), LOS (r = −0.56; p = 0.04) and Patient Health Questionnaire 9 scores (r = −0.72; p = 0.01). Six participants (55%) returned to employment.

Conclusion

Clinically important improvements in biopsychosocial outcomes related to physical function and social factors were observed. Limitations in mental aspects of HRQOL were present at 6 months and some reported mild depressive symptoms.

Clinical implications

Intensive care unit survivors with a history of prolonged MV duration and hospital LOS who exhibit limitations in mental HRQOL, and signs of depressive symptoms should be referred to a psychologist for evaluation.

Factors associated with life-sustaining treatment restriction in a general intensive care unit

Purpose

Few previous studies have investigated associations between clinical variables available after 24 hours in the intensive care unit (ICU), including the Charlson Comorbidity Index (CCI), and decisions to restrict life-sustaining treatment. The aim of this study was to identify factors associated with the life-sustaining treatment restriction and to explore if CCI contributes to explaining decisions to restrict life-sustaining treatment in the ICU at a university hospital in Norway from 2007 to 2009.

Methods

Patients’ Simplified Acute Physiology Score II (SAPS II), age, sex, type of admission, and length of hospital stay prior to being admitted to the unit were recorded. We retrospectively registered the CCI for all patients based on the medical records prior to the index stay. A multivariable logistic regression analysis was used to assess factors associated with treatment restriction during the ICU stay.

Results

We included 936 patients, comprising 685 (73%) medical, 204 (22%) unscheduled and 47 (5%) scheduled surgical patients. Treatment restriction was experienced by 241 (26%) patients during their ICU stay. The variables that were significantly associated with treatment restriction in multivariable analysis were older age (odds ratio [OR] = 1.48 per 10 years, 95% confidence interval [CI] = 1.28–1.72 per 10 years), higher SAPS II (OR = 1.05, 95% CI = 1.04–1.07) and CCI values relative to the reference of CCI = 0: CCI = 2 (OR = 2.08, 95% CI = 1.20–3.61) and CCI≥3 (OR = 2.72, 95% CI = 1.65–4.47).

Conclusions

In multivariable analysis, older age, greater illness severity after 24 h in the ICU and greater comorbidity at hospital admission were independently associated with subsequent life-sustaining treatment restriction. The CCI score contributed additional information independent of the SAPS II illness severity rating.

Optimizing Protein Intake and Nitrogen Balance (OPINiB) in Adult Critically Ill Patients: A Study Protocol for a Randomized Controlled Trial

BACKGROUND

Adequate nutrition of critically ill patients plays a key role in the modulation of metabolic response to stress.

OBJECTIVE

This paper presents the development of a protocol for a randomized controlled trial (RCT) aimed at comparing clinical outcomes of patients in the intensive care unit (ICU) administered with standard and protein-fortified diet. Together with the RCT study protocol, the results of the observational analysis conducted to assess the feasibility of the RCT are presented.

METHODS

An RCT on adult patients admitted to ICU and undergoing mechanical ventilation in the absence of renal or hepatic failure will be conducted. Patients enrolled will be randomized with an allocation rate of 1:1 at standard diet versus protein-fortified diet. The estimated sample size is 19 per arm, for a total of 38 patients to be randomized.

RESULTS

Enrollment began in January 2017. In the feasibility study, 14 patients were enrolled. Protein administration increased significantly (P<.001) over time but was significantly lower compared to that recommended (P<.001). Blood urea nitrogen significantly increased (P<.03) over the period of observation. Such increased catabolism resulted in negative cumulative nitrogen balance (NB) in all patients, and some patients presented with a more negative NB compared to the others.

CONCLUSIONS

Results of the feasibility study clearly confirmed that protein provision in ICU patients is below that recommended and that this results in impaired NB. The emerging of an interindividual variability in NB will be further analyzed in the RCT.

TRIAL REGISTRATION

ClinicalTrials.gov NCT02990065; https://clinicaltrials.gov/ct2/show/NCT02990065 (Archived by WebCite at http://www.webcitation.org/6prsqZdRM).

Brainstem response patterns in deeply-sedated critically-ill patients predict 28-day mortality

Background and purpose

Deep sedation is associated with acute brain dysfunction and increased mortality. We had previously shown that early-assessed brainstem reflexes may predict outcome in deeply sedated patients. The primary objective was to determine whether patterns of brainstem reflexes might predict mortality in deeply sedated patients. The secondary objective was to generate a score predicting mortality in these patients.

Methods

Observational prospective multicenter cohort study of 148 non-brain injured deeply sedated patients, defined by a Richmond Assessment sedation Scale (RASS) <-3. Brainstem reflexes and Glasgow Coma Scale were assessed within 24 hours of sedation and categorized using latent class analysis. The Full Outline Of Unresponsiveness score (FOUR) was also assessed. Primary outcome measure was 28-day mortality. A “Brainstem Responses Assessment Sedation Score” (BRASS) was generated.

Results

Two distinct sub-phenotypes referred as homogeneous and heterogeneous brainstem reactivity were identified (accounting for respectively 54.6% and 45.4% of patients). Homogeneous brainstem reactivity was characterized by preserved reactivity to nociceptive stimuli and a partial and topographically homogenous depression of brainstem reflexes. Heterogeneous brainstem reactivity was characterized by a loss of reactivity to nociceptive stimuli associated with heterogeneous brainstem reflexes depression. Heterogeneous sub-phenotype was a predictor of increased risk of 28-day mortality after adjustment to Simplified Acute Physiology Score-II (SAPS-II) and RASS (Odds Ratio [95% confidence interval] = 6.44 [2.63–15.8]; p<0.0001) or Sequential Organ Failure Assessment (SOFA) and RASS (OR [95%CI] = 5.02 [2.01–12.5]; p = 0.0005). The BRASS (and marginally the FOUR) predicted 28-day mortality (c-index [95%CI] = 0.69 [0.54–0.84] and 0.65 [0.49–0.80] respectively).

Conclusion

In this prospective cohort study, around half of all deeply sedated critically ill patients displayed an early particular neurological sub-phenotype predicting 28-day mortality, which may reflect a dysfunction of the brainstem.

Patients Admitted to Three Spanish Intensive Care Units for Poisoning: Type of Poisoning, Mortality, and Functioning of Prognostic Scores Commonly Used

Objectives. To evaluate the gravity and mortality of those patients admitted to the intensive care unit for poisoning. Also, the applicability and predicted capacity of prognostic scales most frequently used in ICU must be evaluated. Methods. Multicentre study between 2008 and 2013 on all patients admitted for poisoning. Results. The results are from 119 patients. The causes of poisoning were medication, 92 patients (77.3%), caustics, 11 (9.2%), and alcohol, 20 (16,8%). 78.3% attempted suicides. Mean age was 44.42 ± 13.85 years. 72.5% had a Glasgow Coma Scale (GCS) ≤8 points. The ICU mortality was 5.9% and the hospital mortality was 6.7%. The mortality from caustic poisoning was 54.5%, and it was 1.9% for noncaustic poisoning (p < 0.001). After adjusting for SAPS-3 (OR: 1.19 (1.02–1.39)) the mortality of patients who had ingested caustics was far higher than the rest (OR: 560.34 (11.64–26973.83)). There was considerable discrepancy between mortality predicted by SAPS-3 (26.8%) and observed (6.7%) (Hosmer-Lemeshow test: H = 35.10; p < 0.001). The APACHE-II (7,57%) and APACHE-III (8,15%) were no discrepancies. Conclusions. Admission to ICU for poisoning is rare in our country. Medication is the most frequent cause, but mortality of caustic poisoning is higher. APACHE-II and APACHE-III provide adequate predictions about mortality, while SAPS-3 tends to overestimate.

Patient length of stay and mortality prediction: A survey

Over the past few years, there has been increased interest in data mining and machine learning methods to improve hospital performance, in particular hospitals want to improve their intensive care unit statistics by reducing the number of patients dying inside the intensive care unit. Research has focused on prediction of measurable outcomes, including risk of complications, mortality and length of hospital stay. The length of stay is an important metric both for healthcare providers and patients, influenced by numerous factors. In particular, the length of stay in critical care is of great significance, both to patient experience and the cost of care, and is influenced by factors specific to the highly complex environment of the intensive care unit. The length of stay is often used as a surrogate for other outcomes, where those outcomes cannot be measured; for example as a surrogate for hospital or intensive care unit mortality. The length of stay is also a parameter, which has been used to identify the severity of illnesses and healthcare resource utilisation. This paper examines a range of length of stay and mortality prediction applications in acute medicine and the critical care unit. It also focuses on the methods of analysing length of stay and mortality prediction. Moreover, the paper provides a classification and evaluation for the analytical methods of the length of stay and mortality prediction associated with a grouping of relevant research papers published in the years 1984 to 2016 related to the domain of survival analysis. In addition, the paper highlights some of the gaps and challenges of the domain.

Risk factors for relapse or persistence of bacteraemia caused by Enterobacter spp.: a case-control study

Background

Enterobacter spp. possess chromosomal AmpC beta-lactamases that may be expressed at high levels. Previous studies have demonstrated a risk of relapsed bacteraemia following therapy with third generation cephalosporins (3GCs). What additional factors predict microbiological failure in Enterobacter bacteraemia is unclear. We aimed to determine factors associated with microbiological failure in Enterobacter bacteraemia.

Methods

We retrospectively identified cases of bacteraemia caused by Enterobacter spp. occurring in four hospitals. Using a case–control design, we determined clinical risk factors for persistence or relapse defined as repeated positive blood cultures collected between 72 hours and up to 28 days post initial positive blood culture.

Results

During the study period a total of 922 bacteraemia events caused by Enterobacter spp. in adults were identified. The overall risk of relapsed or persisting bacteraemia at 28 days was low (31 of 922, 3.4%), with only 2 patients experiencing emergent resistance to 3GCs. A total of 159 patients were included in the case–control study. Using multivariate logistic regression, independent predictors for relapse were a line-associated source of infection (OR 3.87; 95% CI 1.56-9.60, p = 0.004) and the presence of immunosuppression (OR 2.70; 95% CI 1.14-6.44, p = 0.02). On univariate analysis definitive therapy with a broad-spectrum beta-lactam-beta-lactamase inhibitor (BLBLI, e.g. piperacillin-tazobactam) was not associated with relapse (OR 1.83; 95% CI 0.64-5.21, p = 0.26) although the proportion of patients receiving a BLBLI as definitive therapy was relatively small (21/159, 13.2%).

Conclusions

The risk of relapsed or persistent Enterobacter bacteraemia appears to be low in Australia. A line-associated source of infection and immunocompromise were significant independent predictors for relapse. Larger, preferably randomized, studies are needed to address whether BLBLIs represent an effective carbapenem-sparing option for Enterobacter bacteraemia.

Electronic supplementary material

The online version of this article (doi:10.1186/s13756-017-0177-0) contains supplementary material, which is available to authorized users.

Predictive Performance of the Simplified Acute Physiology Score (SAPS) II and the Initial Sequential Organ Failure Assessment (SOFA) Score in Acutely Ill Intensive Care Patients: Post-Hoc Analyses of the SUP-ICU Inception Cohort Study

Purpose

Severity scores including the Simplified Acute Physiology Score (SAPS) II and the Sequential Organ Failure Assessment (SOFA) score are used in intensive care units (ICUs) to assess disease severity, predict mortality and in research. We aimed to assess the predictive performance of SAPS II and the initial SOFA score for in-hospital and 90-day mortality in a contemporary international cohort.

Methods

This was a post-hoc study of the Stress Ulcer Prophylaxis in the Intensive Care Unit (SUP-ICU) inception cohort study, which included acutely ill adults from ICUs across 11 countries (n = 1034). We compared the discrimination of SAPS II and initial SOFA scores, compared the discrimination of SAPS II in our cohort with the original cohort, assessed the calibration of SAPS II customised to our cohort, and compared the discrimination for 90-day mortality vs. in-hospital mortality for both scores. Discrimination was evaluated using areas under the receiver operating characteristics curves (AUROC). Calibration was evaluated using Hosmer-Lemeshow’s goodness-of-fit Ĉ-statistic.

Results

AUROC for in-hospital mortality was 0.80 (95% confidence interval (CI) 0.77–0.83) for SAPS II and 0.73 (95% CI 0.69–0.76) for initial SOFA score (P<0.001 for the comparison). Calibration of the customised SAPS II for predicting in-hospital mortality was adequate (P = 0.60). Discrimination of SAPS II was reduced compared with the original SAPS II validation sample (AUROC 0.80 vs. 0.86; P = 0.001). AUROC for 90-day mortality was 0.79 (95% CI 0.76–0.82; P = 0.74 for comparison with in-hospital mortality) for SAPS II and 0.71 (95% CI 0.68–0.75; P = 0.66 for comparison with in-hospital mortality) for the initial SOFA score.

Conclusions

The predictive performance of SAPS II was similar for in-hospital and 90-day mortality and superior to that of the initial SOFA score, but SAPS II’s performance has decreased over time. Use of a contemporary severity score with improved predictive performance may be of value.

Construction and internal validation of a new mortality risk score for patients admitted to the intensive care unit

BACKGROUND

The existing models to predict mortality in intensive care units (ICU) present difficulties in clinical practice.

OBJECTIVES

The aim of this study was to develop and internally validate a points system to predict mortality in the ICU, which can be applied instantly and with high discriminating power.

METHODS

This cohort study comprised all patients admitted to the ICU in a Spanish region between January 2013 and April 2014, followed from admission to death or discharge (N=1113). Primary variable: ICU mortality. Secondary variables at admission: gender, Fried criteria for frailty, function scale, medical admission, cardiac arrest, cardiology admission, sepsis, mechanical ventilation, inotropic support, age, frailty index and clinical frailty scale. The sample was divided randomly into two groups (80% and 20%): construction (n=844) and internal validation (n=269). Construction: A logistic regression model was implemented and adapted to the points system.

VALIDATION

the area under the ROC curve (AUC) of the model was calculated and the risk quintiles were created to determine whether differences existed between observed and expected deaths.

RESULTS

The points system included: function scale, medical admission, cardiology admission, sepsis, mechanical ventilation and inotropic support. The validation showed: (i) AUC=0.95 (95% CI: 0.91-0.99, p<.001); (ii) No differences between observed and expected deaths (p=.799).

CONCLUSIONS

A predictive model of mortality in the ICU has been constructed and internally validated. This model improves on the previous models through its simplicity, its discriminating power and free use. External validation studies are needed in other geographical areas.

Impact of socioeconomic status on mortality and unplanned readmission in septic intensive care unit patients

BACKGROUND

Little is known about the potential association between socioeconomic status (SES) and prognosis after sepsis. We analysed how SES impacted mortality and readmission in septic patients treated at the intensive care unit (ICU) of a university hospital.

METHODS

We performed a cohort study including all adult patients admitted to a general tertiary ICU with severe sepsis or septic shock during 2008-2010. Data on SES (educational level, personal income, and cohabitation), comorbidity, readmissions, and mortality were obtained from public registries. We used Cox regression analysis to examine the impact of SES on 30- and 180-day mortality and on first unplanned readmission within 180 days after hospital discharge.

RESULTS

A total of 387 patients were included of whom 111 (29%) died within 30 days after ICU admission, and 55 (20%) died within 180 days after hospital discharge. Adjusted for sex, comorbidity and SAPS II, patients with low income had a substantially greater risk of dying within 30 days of admission compared to those with high income (35.7% vs. 23.3%; adjusted hazard ratio (HR) 1.99; 95% confidence interval (CI) 1.24-3.21), and tended to show higher 180-day mortality (25.0% vs. 15.5%; adjusted HR 1.72; 95% CI 0.86-3.45). Among patients discharged from hospital, 125 (45%) were readmitted within 180 days. Patients with low education and low income showed a tendency towards early readmission.

CONCLUSIONS

Among septic ICU patients, low income was significantly associated with increased 30-day mortality. There was a trend towards earlier readmission among surviving patients with low educational level and personal income.

PIRO concept: staging of sepsis

Introduction:

Sepsis is common presenting illness to the emergency services and one of the leading causes of hospital mortality. Researchers and clinicians have realized that the systemic inflammatory response syndrome concept for defining sepsis is less useful and lacks specificity. The predisposition, infection (or insult), response and organ dysfunction (PIRO) staging of sepsis similar to malignant diseases (TNM staging) might give better information.

Materials and Methods:

A prospective observational study was conducted in emergency medical services attached to medicine department of a tertiary care hospital in Northern India. Patients with age 18 years or more with proven sepsis were included in the first 24 hours of the diagnosis. Two hundred patients were recruited. Multivariate logistic regression analysis was done to assess the factors that predicted in-hospital mortality.

Results:

Two hundred patients with proven sepsis, admitted to the emergency medical services were analysed. Male preponderance was noted (M: F ratio = 1.6:1). Mean age of study cohort was 50.50 ± 16.30 years. Out of 200 patients, 116 (58%) had in-hospital mortality. In multivariate logistic regression analysis, the factors independently associated with in-hospital mortality for predisposition component of PIRO staging were age >70 years, chronic obstructive pulmonary disease, chronic liver disease, cancer and presence of foley's catheter; for infection/insult were pneumonia, urinary tract infection and meningitis/encephalitis; for response variable were tachypnea (respiratory rate >20/minute) and bandemia (band >5%). Organ dysfunction variables associated with hospital mortality were systolic blood pressure <90mm Hg, prolonged activated partial thromboplastin time, raised serum creatinine, partial pressure of oxygen in arterial blood/fraction of inspired oxygen (PaO₂/FiO₂) ratio <300, decreased urine output in first two hours of emergency presentation and Glasgow coma scale ≤9. Each of the components of PIRO had good predictive capability for in-hospital mortality but the total score was more accurate than the individual score and increasing PIRO score was associated with higher in-hospital mortality. The area under receiver operating characteristic curve for cumulative PIRO staging system as a predictor of in-hospital mortality was 0.94.

Conclusion:

This study finds PIRO staging as an important tool to stratify and prognosticate hospitalised patients with sepsis at a tertiary care center. The simplicity of score makes it more practical to be used in busy emergencies as it is based on four easily assessable components.