Predictor Of Intensive Care Unit Mortality Research Articles

The utility of brain natriuretic peptide as a prognosticating marker in critical care patients.

Brain natriuretic peptide (BNP) is an established biomarker of morbidity and mortality in cardiac failure. Data also suggest potential prognostic utility in non-heart failure cohorts. The utility of BNP in predicting intensive care unit (ICU) outcomes has not been well evaluated in a mixed critical care population in the South African (SA) context. To evaluate the ability of BNP to predict ICU mortality in a heterogeneous critical care population in SA. This was a retrospective observational study of 100 patients admitted to a multidisciplinary, closed, intensivist-run ICU in a tertiary academic hospital serving KwaZulu-Natal Province (1 January 2020 - 31 July 2022). Initial BNP was evaluated as a predictor of ICU mortality using univariate and multivariable analyses. There was a statistically significant difference in BNP between survivors and non-survivors in the cohort of patients without heart failure. The median initial BNP in the non-heart failure cohort was 411 (interquartile range (IQR) 116 - 848) ng/L in non-survivors, and 150 (44 - 356) ng/L in survivors (p=0.028). The optimal cut-off for BNP was determined as 366 ng/L. A BNP ≥366 ng/L was an independent predictor of ICU outcome. This study highlights the potential utility of BNP as a predictor of ICU mortality in a heterogeneous ICU population, with the greatest utility in patients without heart failure. Further studies are required to confirm this finding. The study is a retrospective, observational study conducted in multidisciplinary, closed, intensivist-run ICU at a tertiary academic hospital. It showed an elevated BNP is associated with increased ICU mortality, particularly in those without a baseline diagnosis of heart failure. This identifies the need for further prospective studies evaluating BNP as a prognostic marker in non-cardiac critically ill patients, and its utility as an addition in pre-existing ICU outcome prediction scores.

Mortality and Prognostic Factors of Adult Patients with Acute Leukemia Admitted to the ICU: A Systematic Review

Introduction: Acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) are the most aggressive group of hematologic malignancies (Pulte et al, 2020) and will be responsible for approximately 12,700 deaths in the U.S. during 2023(Seiger et al,2023). Improved overall survival (OS) has been observed in the past decades in both diseases, but high mortality in the Intensive Care Unit (ICU) persists, which may prevent ICU admission of those patients. We aimed to review the literature on mortality data and survival predictors of patients with acute leukemia (AL) admitted to the ICU. Methods: Embase, Pubmed and Cochrane databases were searched for articles published from January of 2005 to December of 2020. Due to OS improvement of patients with AML, we only included studies published as of 2005 (Percival et al, 2015). Editorials, case reports and small case series were excluded. Results: Initial search revealed 1,015 manuscripts from database and references, 650 were screened and 115 were fully assessed, for which 24 met eligibility criteria (Figure 1). A total of 13,350 patients were included in the review. Median age was 54.8 years and 57.3% of patients were male. AML represented the majority of diagnoses (98.2%).Two studies included only newly diagnosed patients, 1 study only included patients with relapsed disease and the rest (N=21) included subjects regardless of the AL status. Only 0.7% of patients had a history of hematopoietic stem cell transplantation . 81.9% of patients required mechanical ventilation, 25.3% needed renal replacement therapy, and 65.9% used vasopressors in studies data was available. Median SOFA score in the first 24h of ICU admission was 10.85 . Pooled ICU mortality and hospital mortality were 45.0% and 43.1%, respectively; 58.2% of patients were deceased at 90-day mark (Table 1). The 12 months OS was 37.3%. Higher SOFA score, need for mechanical ventilation, sepsis, vasopressor use, and age greater than 60 years were the most common predictors of ICU, hospital and 90-day mortality. Relapsed or refractory status of underlying AL was the most significant risk factor for long-term mortality. Conclusions: Our systemic review showed that ICU mortality in patients with AL was 45.0%, which represents a survival improvement compared to prior studies on patients with hematological malignancies, where ICU mortality could be as high as 69% (Ashkenazi et al, 1986). Predictors of ICU mortality were closely related to the severity of acute illness rather than malignancy status. We acknowledge that our patient population was mostly comprised of AML cases and the majority of the original studies included a small number of patients with AL with exception of one large study (Halpern et al, 2017).

A cross-modal clinical prediction system for intensive care unit patient outcome

Electronic Health Records (EHRs) are becoming an encyclopedia of clinical concepts, encompassing diagnoses, medications and clinical descriptions to describe patients’ health conditions. The abundance of EHR data presents unprecedented opportunities for leveraging machine learning in clinical predictive tasks. However, the complex process of acquiring healthcare data in some settings like Intensive Care Unit (ICU) gives rise to various challenges in mining EHR, such as modality diversity, heterogeneity, data silo and data quality problems, etc. In this paper, to overcome the data heterogeneity and data quality issues, we propose a Code-Text cross-modal Contrastive Learning (CTCL) framework specifically designed for predicting ICU patient outcomes. This framework leverages both structured (multi-view medical codes) and unstructured (clinical notes) EHR data to extract patient representations. To capture the complicated intra and inter interactions among medical codes, CTCL employs a multi-view graph convolution network to extract code encodings and a cross-view contrastive loss to regularize the model parameters. To enhance patient representations across different modalities, a cross-modal contrastive learning module is introduced. Experimental evaluations on the eICU-CRD database demonstrate that CTCL outperforms state-of-the-art competitors on the ICU mortality prediction task and in-hospital readmission prediction task. Furthermore, ablation studies and discussions demonstrate the effectiveness of each module of the framework.

Increased Levels of Galectin-3 in Critical COVID-19.

Severe COVID-19 is related to hyperinflammation and multiple organ injury, including respiratory failure, thus requiring intensive care unit (ICU) admission. Galectin-3, a carbohydrate-binding protein exhibiting pleiotropic effects, has been previously recognized to participate in inflammation, the immune response to infections and fibrosis. The aim of this study was to evaluate the relationship between galectin-3 and the clinical severity of COVID-19, as well as assess the prognostic accuracy of galectin-3 for the probability of ICU mortality. The study included 235 COVID-19 patients with active disease, treated in two different Greek hospitals in total. Our results showed that median galectin-3 serum levels on admission were significantly increased in critical COVID-19 patients (7.2 ng/mL), as compared to the median levels of patients with less severe disease (2.9 ng/mL, p = 0.003). Galectin-3 levels of the non-survivors hospitalized in the ICU were significantly higher than those of the survivors (median 9.1 ng/mL versus 5.8 ng/mL, p = 0.001). The prognostic accuracy of galectin-3 for the probability of ICU mortality was studied with a receiver operating characteristic (ROC) curve and a multivariate analysis further demonstrated that galectin-3 concentration at hospital admission could be assumed as an independent risk factor associated with ICU mortality. Our results were validated with galectin-3 measurements in a second patient cohort from a different Greek university hospital. Our results, apart from strongly confirming and advancing previous knowledge with two patient cohorts, explore the possibility of predicting ICU mortality, which could provide useful information to clinicians. Therefore, galectin-3 seems to establish its involvement in the prognosis of hospitalized COVID-19 patients, suggesting that it could serve as a promising biomarker in critical COVID-19.

A new efficient ALignment-driven Neural Network for Mortality Prediction from Irregular Multivariate Time Series data

The irregularity of the time interval between observations in and across the stream is a key factor that leads to a drop in performance when classical machine learning or deep learning models are used for a downstream task requiring multivariate time series. Indeed, irregular multivariate time series not only increase the rate of missing values but also lead to data sparsity, which consequently makes the data almost unleverageable and/or ineffective for models. To tackle this scorching challenge, most of the pioneering approaches apply imputation or interpolation in their core, which might lead to embedding data with noise. To especially address this irregular multivariate time series issue, we introduce, in this paper, a new deep neural network model called ALignment-driven Neural Network. The innovative idea of our model is to transform the irregular multivariate time series into pseudo-aligned (or pseudo-regular) latent values. The latter are shown as a matrix, where the coefficients are the latent values of each feature at user-defined reference time points that are evenly spaced. They are obtained through a duplication process driven by an exponential decay mechanism. The obtained output is then passed to a Recurrent Neural Network model, which is undoubtedly the must-use model for regular time series data. To show that our model added value, we looked at the Intensive Care Unit mortality prediction task. In this unit, the physiological measurements used to make decisions have a problem with time irregularity. Leveraging the publicly available MIMIC-III, we compare the performance of our model to that of flagship models. In addition, we also performed extensive ablation studies to highlight the importance of specific components in our model. Interestingly enough, whenever data is collected 24 and 48 h after a patient’s admission, we outperform our pioneering competitors, i.e., +1.1% and +1.5% for the AUC score, +2.3% and +2.4% for the AUPRC score and +0.6% and +1.7% for the F1-score.

Serum Cholinesterase, C-reactive Protein, Interleukin 6, and Procalcitonin Levels as Predictors of Mortality in Patients in the Intensive Care Unit.

The prognostic utility of inflammatory markers in survival has been suggested in patients with cancer; however, evidence on their prognostic value in severely ill patients is very limited. We aimed to explore the prognostic value of cholinesterase (ChE), C-reactive protein (CRP), interleukin-6 (IL-6), and procalcitonin (PCT) in predicting mortality in patients from the intensive care unit (ICU). Serum levels of ChE, CRP, IL-6 and PCT were measured in ICU patients from December 13th, 2019 to June 28th, 2022. We assessed the predictive power of ChE, CRP, IL-6, and PCT using the receiver operating characteristic (ROC) curves. Furthermore, we evaluated their diagnostic accuracy by comparing the areas under the ROC curve (AUCs) along with their corresponding 95% confidence intervals (CIs). The cut-off values were determined to dichotomise these biomarkers, which were then included in multivariable logistic regression models to examine their relationship with ICU mortality. Among 253 ICU patients included in the study, 66 (26%) died during the ICU stay. The AUCs to predict ICU mortality were 0.643 (95% CI, 0.566-0.719), 0.648 (95% CI, 0.633-0.735), 0.643 (95% CI, 0.563-0.723) and 0.735 (95% CI, 0.664-0.807) for ChE, CRP, IL-6 and PCT, respectively. After adjusting for age, sex and disease severity, lower ChE level (<3.668 × 103 U L-1) and higher levels of CRP (>10.546 mg dL-1), IL-6 (>986.245 pg mL-1) and PCT (>0.505 μg L-1) were associated with higher mortality risk, with odd ratios of 2.70 (95% CI, 1.32-5.54), 4.99 (95% CI, 2.41-10.38), 3.24 (95% CI, 1.54-6.78) and 3.67 (95% CI, 1.45-9.95), respectively. ChE, CRP, IL-6 and PCT were independent ICU mortality risk factors in severely ill patients. Elevated PCT levels exhibited better predictive value than the other three biomarkers that were evaluated.

Discriminative performance of pancreatic stone protein in predicting ICU mortality and infection severity in adult patients with infection: a systematic review and individual patient level meta-analysis

BackgroundSeveral studies suggested pancreatic stone protein (PSP) as a promising biomarker to predict mortality among patients with severe infection. The objective of the study was to evaluate the performance of PSP in predicting intensive care unit (ICU) mortality and infection severity among critically ill adults admitted to the hospital for infection.MethodsA systematic search across Cochrane Central Register of Controlled Trials and MEDLINE databases (1966 to February 2022) for studies on PSP published in English using ‘pancreatic stone protein’, ‘PSP’, ‘regenerative protein’, ‘lithostatin’ combined with ‘infection’ and ‘sepsis’ found 46 records. The search was restricted to the five trials that measured PSP using the enzyme-linked immunosorbent assay technique (ELISA). We used Bayesian hierarchical regression models for pooled estimates and to predict mortality or disease severity using PSP, C-Reactive Protein (CRP) and procalcitonin (PCT) as main predictor. We used statistical discriminative measures, such as the area under the receiver operating characteristic curve (AUC) and classification plots.ResultsAmong the 678 patients included, the pooled ICU mortality was 17.8% (95% prediction interval 4.1% to 54.6%) with a between-study heterogeneity (I-squared 87%). PSP was strongly associated with ICU mortality (OR = 2.7, 95% credible interval (CrI) [1.3–6.0] per one standard deviation increase; age, gender and sepsis severity adjusted OR = 1.5, 95% CrI [0.98–2.8]). The AUC was 0.69 for PSP 95% confidence interval (CI) [0.64–0.74], 0.61 [0.56–0.66] for PCT and 0.52 [0.47–0.57] for CRP. The sensitivity was 0.96, 0.52, 0.30 for risk thresholds 0.1, 0.2 and 0.3; respective false positive rate values were 0.84, 0.25, 0.10.ConclusionsWe found that PSP showed a very good discriminative ability for both investigated study endpoints ICU mortality and infection severity; better in comparison to CRP, similar to PCT. Combinations of biomarkers did not improve their predictive ability.

Efficacy of Neutrophil Lymphocyte Stress Factor in Diagnosis of Sepsis and Prediction of Mortality in Septic Patients Admitted to the ICU

Abstract Background Sepsis is a major public health concern. It is among the most common reasons for admission to intensive care units (ICUs) throughout the world. Severe sepsis and septic shock are associated with 30-60% mortality rate which is very high compared to other common diseases, such as myocardial infarction or breast cancer. Sepsis is now defined as a “life-threatening organ dysfunction due to a dysregulated host response to infection”. Aim of the Work to evaluate the efficacy of neutrophil lymphocyte stress factor (NLSF) in the diagnosis of sepsis and prediction of mortality in septic patients admitted to the ICU. Patients and Methods We conducted this prospective observational study on 60 patients after informed consents were taken from patients or their next of kin. We enrolled adult patients admitted to the ICU with the diagnosis of septic shock. We excluded patients younger than 18 years, pregnant females, patients with terminal malignancy, patients with liver cirrhosis child-pugh C, patients with shock of mixed origin and patients with aplasia, immunosuppressive disease or receiving immunosuppressive therapy. Then we classified patients into 2 main groups according to presence of sepsis into; Septic group (30 patients), and Non-Septic group (30 patients). Results from our study concluded that NLSF can be used in diagnosis of sepsis, although it’s the least accurate diagnostic marker among PCT, CRP and NLSF, it provides the cheapest tool for diagnosis of sepsis. It was also found that NLSF can be used in prediction of mortality in septic patient during the period of our study and that it’s the most accurate prognostic marker of mortality among PCT, CRP and NLSF. Conclusion NLSF in combination with other parameters can be used as a cheap easy tool for the early detection of sepsis in intensive care units. NLSF can be used as a prognostic marker in sepsis for prediction of mortality in ICU. Further studies are required to clarify these results on a larger group of patients and for longer periods of time.

Investigating the Utility of the SOFA Score and Creating a Modified SOFA Score for Predicting Mortality in the Intensive Care Units in a Tertiary Hospital in Jordan.

The utility of the Sequential Organ Failure Assessment (SOFA) score in predicting mortality in the intensive care unit (ICU) has been demonstrated before, but serial testing in various settings is required to validate and improve the score. This study examined the utility of the SOFA score in predicting mortality in Jordanian ICU patients and aimed to find a modified score that required fewer laboratory tests. A prospective observational study was conducted at Jordan University Hospital (JUH). All adult patients admitted to JUH ICUs between June and December 2020 were included in the study. SOFA scores were measured daily during the whole ICU stay. A modified SOFA score (mSOFA) was constructed from the available laboratory, clinical, and demographic data. The performance of the SOFA, mSOFA, qSOFA, and SIRS in predicting ICU mortality was assessed using the area under the receiver operating characteristic curve (AUROC). 194 patients were followed up. SOFA score (mean ± SD) at admission was significantly higher in non-survivors (7.5 ± 3.9) compared to survivors (2.4 ± 2.2) and performed the best in predicting ICU mortality (AUROC = 0.8756, 95% CI: 0.8117-0.9395) compared to qSOFA (AUROC = 0.746, 95% CI: 0.655-0.836) and SIRS (AUROC = 0.533, 95% CI: 0.425-0.641). The constructed mSOFA included points for the hepatic and CNS SOFA scores, in addition to one point each for the presence of chronic kidney disease or the use of breathing support; it performed as well as the SOFA score in this cohort or better than the SOFA score in a subgroup of patients with heart disease. SOFA score was a good predictor of mortality in a Jordanian ICU population and better than qSOFA, while SIRS could not predict mortality. Furthermore, the proposed mSOFA score which employed fewer laboratory tests could be used after validation from larger studies.

Association between cumulative fluid balance and clinical outcome in patients with septic shock: A retrospective observational study

Abstract Introduction: Aggressive fluid replacement is done in septic shock to compensate for the decreased stressed volume in view of peripheral vasodilatation. However, excessive fluid administration may lead to fluid overload, edema formation, and organ damage such as renal congestion and acute kidney injury. Hence, this study was conducted to estimate the effect of fluid balance in the 1st week of septic shock on mortality. Materials and Methods: We conducted a retrospective analysis of the use of intravenous fluids and cumulative fluid balance (CFB) during the first 7 days of septic shock in a 25-bed multidisciplinary intensive care unit (ICU) in Northern India at a tertiary care teaching hospital. Results: Based on the CFB, we classified patients as low CFB when CFB is &lt;20 mL/kg and high CFB when it is &gt;20 mL/kg. The 28-day mortality in the low CFB group was 4.7% and in the high CFB group 26.15%, which was statistically significant. With an odds ratio of 7.17, CFB in septic shock was a predictor of ICU mortality (confidence interval of 3.87–13). The length of stay between the two groups was also found to be statistically significant (11.7 ± 3.2 vs. 12.6 ± 3.8; P = 0.02). Conclusions: In patients with septic shock, a higher CFB on day 7 of ICU admission is independently associated with an increased risk of mortality and a longer ICU stay. The following core competencies are addressed in this article: Practice-based learning and improvement, Medical knowledge.

The correlation between electrocardiographic parameters and mortality in non-cardiac ICU patients.

This study aimed to determine the correlation between selected electrocardiogram (ECG) parameters (recorded at admission) and mortality in non-cardiac, non-COVID-19 intensive care unit (ICU) patients, and to determine the sensitivity and specificity of a novel admission ECG score (AD-ECG) for predicting mortality. Additionally, the sensitivity and specificity of the AD-ECG and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores for predicting ICU mortality were compared. Clinical and laboratory data, and ECG parameters were compared between ICU survivors and non-survivors. ECG parameters (the QTc and Tpe intervals, and the Tpe/QT and Tpe/QTc ratios) and pulse pressure at ICU admission (baseline) were used to calculate the AD-ECG score. Cut-off values for ECG parameters, pulse pressure, and AD-ECG and APACHE II scores were calculated. The sensitivity and specificity of the APACHE II and AD-ECG scores were determined. The study included 167 patients. Mortality was higher in the patients with comorbidities, mechanical ventilation, and length of ICU stay (p < 0.05). The QTc and Tpe intervals, and the TPe/QT and TPe/QTc ratios differed significantly between the survivors and non-survivors (p < 0.05). The sensitivity and specificity of the AD-ECG score were similar to those of the APACHE II score. When pulse pressure, and the QTc and Tpe intervals were added to APACHE II, the sensitivity of the APACHE II score increased from 78.9% to 85.5%, and its specificity increased from 75% to 86.8%. A novel admission ECG score (AD-ECG) based on ECG parameters (the QTc and Tpe intervals, and the Tpe/QT and Tpe/QTc ratios) and pulse pressure has similar sensitivity and specificity as the APACHE II score for predicting non-cardiac ICU mortality. Adding pulse pressure, and the QTc and Tpe intervals increases the sensitivity and specificity of the APACHE II score; however, as the present study included non-cardiac patients only, additional larger-scale studies are needed to obtain more precise results.

Bacterial and fungal superinfections are detected at higher frequency in critically ill patients affected by SARS CoV-2 infection than negative patients and are associated to a worse outcome.

Patients with viral infections are at higher risk to acquire bacterial and fungal superinfections associated with a worse prognosis. We explored this critical point in the setting of patients with severe COVID-19 disease. The study included 1911 patients admitted to intensive care unit (ICU) during a 2-year study period (March 2020-March 2022). Of them, 713 (37.3%) were infected with SARS-CoV-2 and 1198 were negative (62.7%). Regression analysis was performed to determine risk factors associated with the presence of bacterial and/or fungal superinfections in SARS-CoV-2 patients and to evaluate predictors of ICU mortality. Of the 713 patients with SARS-CoV-2 infection, 473 (66.3%) had respiratory and/or bloodstream bacterial and/or fungal superinfections, while of the 1198 COVID-19-negative patients, only 369 (30%) showed respiratory and/or bloodstream bacterial and/or fungal superinfections (p < 0.0001). Baseline characteristics of COVID-19 patients included a median age of 66 (interquartile range [IQR], 58-73), a predominance of males (72.7%), and the presence of a BMI higher than 24 (median 26; IQR, 24.5-30.4). Seventy-four percent (527, 73.9%) had one or more comorbidities and 135 (18.9%) of them had received previous antibiotic therapy. Furthermore, most of them (473, 66.3%) exhibited severe radiological pictures and needed invasive mechanical ventilation. Multivariate logistic regression analysis showed that 1 unit increment in BMI rises the risk of bacterial and/or fungal superinfections acquisition by 3% and 1-day increment in ICU stays rises the risk of bacterial and/or fungal superinfections acquisition by 11%. Furthermore, 1-day increment in mechanical ventilation rises the risk of bacterial and/or fungal superinfection acquisition by 2.7 times. Furthermore, patients with both bacterial and fungal infections had a significantly higher mortality rate than patients without superinfections (45.8% vs. 26.2%, p < 0.0001). Therefore, bacterial and fungal superinfections are frequent in COVID-19 patients admitted to ICU and their presence is associated with a worse outcome. This is an important consideration for targeted therapies in critically ill SARS-CoV-2 infected patients to improve their clinical course.

Optimized Risk Score to Predict Mortality in Patients With Cardiogenic Shock in the Cardiac Intensive Care Unit.

Background Mortality prediction in critically ill patients with cardiogenic shock can guide triage and selection of potentially high-risk treatment options. Methods and Results We developed and externally validated a checklist risk score to predict in-hospital mortality among adults admitted to the cardiac intensive care unit with Society for Cardiovascular Angiography & Interventions Shock Stage C or greater cardiogenic shock using 2 real-world data sets and Risk-Calibrated Super-sparse Linear Integer Modeling (RiskSLIM). We compared this model to those developed using conventional penalized logistic regression and published cardiogenic shock and intensive care unit mortality prediction models. There were 8815 patients in our training cohort (in-hospital mortality 13.4%) and 2237 patients in our validation cohort (in-hospital mortality 22.8%), and there were 39 candidate predictor variables. The final risk score (termed BOS,MA2) included maximum blood urea nitrogen ≥25 mg/dL, minimum oxygen saturation <88%, minimum systolic blood pressure <80 mm Hg, use of mechanical ventilation, age ≥60 years, and maximum anion gap ≥14 mmol/L, based on values recorded during the first 24 hours of intensive care unit stay. Predicted in-hospital mortality ranged from 0.5% for a score of 0to 70.2% for a score of 6. The area under the receiver operating curve was 0.83 (0.82-0.84) in training and 0.76 (0.73-0.78) in validation, and the expected calibration error was 0.9% in training and 2.6% in validation. Conclusions Developed using a novel machine learning method and the largest cardiogenic shock cohorts among published models, BOS,MA2 is a simple, clinically interpretable risk score that has improved performance compared with existing cardiogenic-shock risk scores and better calibration than general intensive care unit risk scores.

Comparison between presepsin, procalcitonin, and CRP as biomarkers to diagnose sepsis in critically ill patients.

Mortality associated with sepsis continues to remain high. Early diagnosis and aggressive management may improve outcomes. Biomarkers may help in early diagnosis, but the search for an ideal biomarker continues. Presepsin has been introduced as a new biomarker, however, it still needs validation before its use becomes routine. In this study, we aimed to compare the efficacy of various biomarkers in patients with suspected sepsis. A retrospective analysis of 100 patients with suspected infection, admitted in the medical intensive care unit (ICU) was conducted. Diagnosis of sepsis was made on the basis of the current surviving sepsis guidelines criteria. Out of 100 patients, 70 were diagnosed to have sepsis, and overall ICU mortality was 22%. Overall, C-reactive protein (CRP) was positive in 98, procalcitonin in 75, and presepsin in 64 patients. For diagnosis of sepsis the sensitivity, specificity, and AUC, respectively, for CRP was 98.6%, 3.3%, and 0.725. For procalcitonin (>0.5 ng/ml) it was 87.1%, 53.3%, and 0.776, and for procalcitonin (>1 ng/ml) 70%, 70%, and 0.816, respectively. For presepsin sensitivity, specificity, and AUC, respectively, for diagnosis of sepsis was 77.1%, 66.7%, and 0.734. For ICU mortality, sensitivity and specificity for CRP was 95.5% and 1.3%, for procalcitonin (>0.5) 72.7% and 24.4.%, for procalcitonin (>1) 59.1% and 42.3%, and for presepsin 61.5% and 27.3%, respectively. Inflammatory markers may be raised in a large proportion of ICU patients, even in those without sepsis. Procalcitnonin and presepsin had similar efficacy in diagnosing sepsis. However, none of the three biomarkers studied were accurate in predicting ICU mortality.

Explainable Mortality Prediction Model for Congestive Heart Failure with Nature-Based Feature Selection Method

A mortality prediction model can be a great tool to assist physicians in decision making in the intensive care unit (ICU) in order to ensure optimal allocation of ICU resources according to the patient’s health conditions. The entire world witnessed a severe ICU patient capacity crisis a few years ago during the COVID-19 pandemic. Various widely utilized machine learning (ML) models in this research field can provide poor performance due to a lack of proper feature selection. Despite the fact that nature-based algorithms in other sectors perform well for feature selection, no comparative study on the performance of nature-based algorithms in feature selection has been conducted in the ICU mortality prediction field. Therefore, in this research, a comparison of the performance of ML models with and without feature selection was performed. In addition, explainable artificial intelligence (AI) was used to examine the contribution of features to the decision-making process. Explainable AI focuses on establishing transparency and traceability for statistical black-box machine learning techniques. Explainable AI is essential in the medical industry to foster public confidence and trust in machine learning model predictions. Three nature-based algorithms, namely the flower pollination algorithm (FPA), particle swarm algorithm (PSO), and genetic algorithm (GA), were used in this study. For the classification job, the most widely used and diversified classifiers from the literature were used, including logistic regression (LR), decision tree (DT) classifier, the gradient boosting (GB) algorithm, and the random forest (RF) algorithm. The Medical Information Mart for Intensive Care III (MIMIC-III) dataset was used to collect data on heart failure patients. On the MIMIC-III dataset, it was discovered that feature selection significantly improved the performance of the described ML models. Without applying any feature selection process on the MIMIC-III heart failure patient dataset, the accuracy of the four mentioned ML models, namely LR, DT, RF, and GB was 69.9%, 82.5%, 90.6%, and 91.0%, respectively, whereas with feature selection in combination with the FPA, the accuracy increased to 71.6%, 84.8%, 92.8%, and 91.1%, respectively, for the same dataset. Again, the FPA showed the highest area under the receiver operating characteristic (AUROC) value of 83.0% with the RF algorithm among all other algorithms utilized in this study. Thus, it can be concluded that the use of feature selection with FPA has a profound impact on the outcome of ML models. Shapley additive explanation (SHAP) was used in this study to interpret the ML models. SHAP was used in this study because it offers mathematical assurances for the precision and consistency of explanations. It is trustworthy and suitable for both local and global explanations. It was found that the features that were selected by SHAP as most important were also most common with the features selected by the FPA. Therefore, we hope that this study will help physicians to predict ICU mortality for heart failure patients with a limited number of features and with high accuracy.

Prognostic values of modified NUTRIC score to assess outcomes in critically ill patients admitted to the intensive care units: prospective observational study

PurposeModified Nutrition Risk in the Critically Ill (NUTRIC) score (mNUTRIC score) have been validated as screening tool for quantifying risk of adverse outcome critically ill patients admitted to the intensive care units (ICUs). The aim of this study was to evaluate the prognostic value of mNUTRIC score to assess outcomes in this population.Materials and methodsThis prospective, observational study was conducted on adult patients admitted to the general ICUs of two university affiliated hospital in northwest of Iran. The association between the mNUTRIC score and outcomes was assessed using the univariate and multivariate binary logistic regression. The performance of mNUTRIC score to predict outcomes was assessed using the receiver operating characteristic (ROC)-curve.ResultsIn total 445 ICU patients were enrolled. Based on mNUTRIC score, 62 (13.9%) and 383 (86.1%) individuals were identified at high and low nutritional risk, respectively. The area under the curve (AUC) for predicting ICU mortality, using vasopressor, duration of vasopressor, and mechanical ventilation (MV) duration were (AUC: 0.973, 95% CI: 0.954–0.986, P < 0.001), (AUC: 0.807, 95% CI: 0.767–0.843, P < 0.001), (AUC: 0.726, 95% CI: 0.680–0.769, P < 0.001) and (AUC: 0.710, 95% CI: 0.666–0.752, P < 0.001), respectively.ConclusionsAn excellent and good predictive performance of the mNUTRIC score was found regarding ICU mortality and using vasopressor, respectively. However, this predictive was fair for MV and vasopressor duration and poor for ICU and hospital length of stay.

Dual-core mutual learning between scoring systems and clinical features for ICU mortality prediction

Perpetually improving mortality prediction in intensive care units (ICUs) via the implementation of eHealth evaluation approaches has become a major research hotspot in the field of medical data mining for the purpose of saving lives. Recently, researchers have attempted to achieve improved prediction accuracy by using only deep learning-based techniques. However, some problems remain. (1) Most of the existing methods utilize independent clinical features to predict mortality by eliminating the correlations between the latent features, but this technique may fail to comprehensively capture and evaluate the statuses of patients. (2) Several clinical features are needed to ensure strong prediction accuracy, but most methods only use static features that are not extendable. (3) An effective practical framework that unifies traditional ICU scoring systems and state-of-the-art deep learning methods to predict mortality is also lacking. (4) Moreover, the interpretability of existing deep learning-based methods needs to be further improved. Therefore, we propose a novel dual-core mutual learning framework (DMLF) between ICU scoring systems and clinical features for mortality prediction. In particular, we mutually utilize sequential organ failure assessment (SOFA) scores and clinical measurement features to learn a unified model for enhancing the accuracy and interpretability of our DMLF. Experiments conducted on five real-world disease datasets show that the DMLF achieves significantly better prediction accuracy and area under the receiver operating characteristic curve (AUROC) values than six baselines and four state-of-the-art methods. Moreover, clinicians utilize a familiarized SOFA system to conduct mortality prediction and achieve increased interpretability, which benefits the adoption of the proposed framework in real clinical scenarios.

Sequential Organ Failure Assessment (SOFA) Score for predicting mortality in patients with sepsis in Vietnamese intensive care units: a multicentre, cross-sectional study

ObjectivesTo compare the accuracy of the Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation II (APACHE II) Scores in predicting mortality among intensive care unit (ICU)...

Characteristics and outcomes of patients with chronic obstructive pulmonary disease admitted to the intensive care unit due to acute hypercapnic respiratory failure.

The study aimed to describe the clinical course, outcomes, and prognostic factors of chronic obstructive pulmonary disease (COPD) patients with acute hypercapnic respiratory failure. This retrospective study involved patients with acute hypercapnic respiratory failure due to COPD of any cause admitted to the intensive care unit (ICU) for non-invasive or invasive mechanical ventilation (IMV) support between December 2015 and February 2020. One hundred patients were evaluated. The main causes of acute hypercapnic respiratory failure were bronchitis, pneumonia, and heart failure. The patients' mean Acute Physiology and Chronic Health Evaluation (APACHE) II score was 23.0±7.2, and their IMV rate was 43%. ICU, in-hospital, and 90-day mortality rates were 21%, 29%, and 39%, respectively. Non-survivors had more pneumonia, shock within the first 24 hours of admission, IMV, vasopressor use, and renal replacement therapy, along with higher APACHE II scores, lower admission albumin levels and PaO2/ FiO2 ratios, and longer ICU and hospital stays than survivors. Logistic regression analysis identified APACHE II score (odds ratio [OR], 1.157; 95% confidence interval [CI], 1.017-1.317; P=0.026), admission PaO2/FiO2 ratio (OR, 0.989; 95% CI, 0.978-0.999; P=0.046), and vasopressor use (OR, 8.827; 95% CI, 1.650-47.215; P=0.011) as predictors of ICU mortality. APACHE II score (OR, 1.099; 95% CI, 1.021-1.182; P=0.011) and admission albumin level (OR, 0.169; 95% CI, 0.056-0.514; P=0.002) emerged as predictors of 90-day mortality. APACHE II scores, the PaO2/FiO2 ratio, vasopressor use, and albumin levels are significant short-term mortality predictors in severely ill COPD patients with acute hypercapnic respiratory failure.

Pulmonary Infections and Outcomes in AIDS Patients with Respiratory Failure: A 10-Year Retrospective Review.

Respiratory failure in acquired immunodeficiency syndrome (AIDS) patients was the leading cause of intensive care unit (ICU) admission in our center. We aimed to describe the pulmonary infections and outcomes for respiratory failure in AIDS patients. A retrospective study was conducted on AIDS adult patients with respiratory failure who were admitted to the ICU in Beijing Ditan hospital, China, from January 2012 to December 2021. We investigated pulmonary infections complicated by respiratory failure in AIDS patients. The primary outcome was ICU mortality, and a comparison between survivors and nonsurvivors was performed. Multiple logistic regression analysis was used to identify predictors of ICU mortality. The Kaplan-Meier curve and Log rank test were used for survival analysis. A total of 231 AIDS patients were admitted to ICU with respiratory failure over a 10-year period with a male predominance (95.7%). Pneumocystis jirovecii pneumonia was the main etiology of pulmonary infections (80.1%). The ICU mortality was 32.9%. In multivariate analysis, ICU mortality was independently associated with invasive mechanical ventilation (IMV) [odds ratio (OR), 27.910; 95% confidence interval (CI, 8.392-92.818; p = 0.000) and the time before ICU admission (OR, 0.959; 95% CI, 0.920-0.999; p = 0.046). In the survival analysis, patients with IMV and later admission to ICU had a higher probability of mortality. Pneumocystis jirovecii pneumonia was the primary etiology for respiratory failure in AIDS patients admitted to the ICU. Respiratory failure remains a severe illness with high mortality, and ICU mortality was negatively associated with IMV and later admission to ICU.