Predictors of Intensive Care Unit Admission in Red Code Patients in the Emergency Department: A Single-Center Retrospective Observational Study

ObjectivesThis study aimed to explore the value of the Charlson comorbidity index (CCI) in predicting ICU admission in patients with aortic aneurysm (AA).MethodsThe clinical data of patients were obtained from the Medical Information Mart for Intensive Care-IV database. The association between CCI and ICU admission was explored by restricted cubic spline (RCS), threshold effect analysis, generalized linear model, logistic regression, interaction, and mediation analyses. Its clinical value was evaluated by decision curve analysis (DCA), receiver operating characteristic curve (ROC), DeLong's test, and net reclassification index (NRI) analyses.ResultsThe ICU admission was significantly associated with the thoracic AA (TAA), unruptured status, and surgery status. Therefore, 288 candidate patients with unruptured TAA who received surgery were enrolled in the further analysis. We found that CCI was independently associated with the ICU admission of candidates (P = 0.005). Further, their nonlinear relationship was observed (adjusted P = 0.008), and a significant turning point of 6 was identified. The CCI had a favorable performance in predicting ICU admission (area under curve = 0.728) and achieved a better clinical net benefit. New models based on CCI significantly improved the accuracy of prediction. Besides the importance of CCI in ICU admission, CCI also exerted important interaction effect (rather than mediating effects) on the association of other variables (such as age and blood variables) with ICU admission requirements (all P < 0.05).ConclusionsThe CCI is an important predictor of ICU admission after surgery in patients with unruptured TAA.

On improving assessment of in-hospital mortality and ICU admission in community-acquired pneumonia patients using the eCURB

CAP is the most common cause of death in infectious diseases in developing countries, while also an important cause of death and morbidity in developed countries. In recent years, CURB-65 (or CRB-65) and pneumonia severity index (PSI) scoring systems have been widely used in the prognosis scoring system of CAP. However, each of them has some shortcomings in predicting ICU admission in CAP patients. The aim of this study is to analyze serum inflammatory biomarkers combined age to established a new prediction model in predicting ICU admission in CAP patients. This is a retrospective study. The enrolled CAP patients received serum inflammatory biomarker tests, including procalcitonin (PCT), white blood cell count (WBC), hypersensitive C-reactive protein (hs-CRP), and erythrocyte sedimentation rate (ESR). Body temperature and age were also recorded. The main outcome measures were ICU admission. Univariate analysis and binary logistic regression analysis were used to explore the in-dependent risk factors which could be components of a new predicting model for ICU admission in CAP patients. Receiver operating characteristic curves (ROC) were used to evaluate the sensitivity and specificity of the new model, which consisted of the combination of all independent risk factors in predicting the main outcomes. Initially, 246 CAP patients were admitted to general wards, 61 of whom were subsequently transferred to ICU (61/246). Age, PCT, WBC, and hs-CRP were independent risk factors for subsequent admission to ICU for CAP patients in general wards. The AUC of the ROC curve of new prediction model (the joint model consists of age, PCT, WBC, and hs-CRP) was 0.93 (95% CI 0.85 - 0.96), the sensitivity and specificity were 85.2% and 88.1%, respectively. Serum inflammatory biomarkers combined age have high specificity and sensitivity in predicting ICU admission in adult CAP patients.

BackgroundThe demand for inpatient medical services increases during influenza season. A scoring system capable of identifying influenza patients at low risk death or ICU admission could help clinicians make hospital admission decisions.MethodsHospitalized patients with laboratory confirmed influenza were identified over 3 influenza seasons at 25 Ontario hospitals. Each patient was assigned a score for 6 pneumonia severity and 2 sepsis scores using the first data available following their registration in the emergency room. In-hospital mortality and ICU admission were the outcomes. Score performance was assessed using the area under the receiver operating characteristic curve (AUC) and the sensitivity and specificity for identifying low risk patients (risk of outcome <5%).ResultsThe cohort consisted of 607 adult patients. Mean age was 76 years, 12% of patients died (71/607) and 9% required ICU care (55/607). None of the scores examined demonstrated good discriminatory ability (AUC≥0.80). The Pneumonia Severity Index (AUC 0.78, 95% CI 0.72–0.83) and the Mortality in Emergency Department Sepsis score (AUC 0.77, 95% 0.71–0.83) demonstrated fair predictive ability (AUC≥0.70) for in-hospital mortality. The best predictor of ICU admission was SMART-COP (AUC 0.73, 95% CI 0.67–0.79). All other scores were poor predictors (AUC <0.70) of either outcome. If patients classified as low risk for in-hospital mortality using the PSI were discharged, 35% of admissions would have been avoided.ConclusionsNone of the scores studied were good predictors of in-hospital mortality or ICU admission. The PSI and MEDS score were fair predictors of death and if these results are validated, their use could reduce influenza admission rates significantly.

Introduction: ICU admission and prolonged length-of-stay (PLOS) is associated with increased resources need, healthcare cost and poorer outcomes for the patients. There is a paucity of predictive parameters to assess ICU admission and PLOS in non-variceal upper gastrointestinal bleeding (NVUGIB) patients taking dual antiplatelet therapy (DAPT). Methods: A retrospective chart review was done for all consecutive adult patients between 2015-2020 who presented with NVUGIB on DAPT (aspirin, clopidogrel; aspirin, ticagrelor; aspirin, prasugrel) to Allegheny Health Network and underwent endoscopy. Patients who had DAPT held as per ASGE guidelines were excluded from the study. An exploratory analysis was done to assess predictive ability of four GI bleeding risk scores [Glasgow Blatchford score (GBS), AIMS65, qSOFA and Rockall] to predict ICU admission and prolonged Length of Stay (PLOS). PLOS was defined as ≥5 days based on institutional experience in NVUGIB patients. Logistic regression was used to identify variables independently associated with ICU admission and PLOS and estimate their odds ratios and 95% CI. P value < 0.05 was considered as the level of significance. Results: 75 patients (M: F 53:22), mean age±S.D. 70.1±11.2 years, meeting inclusion criteria were included for final analysis. 22/75 (29%) patients had ICU admission. AIMS65 and qSOFA predicted ICU admission (Table 1A). Sensitivity, specificity, PPV, NPV and accuracy of AIMS65 and qSOFA for ICU admission has been shown in Table 1B. Median LOS was 6; IQR=8 (range 2-43) days. 47/75 (63%) patients had PLOS. GBS and AIMS65 predicted PLOS (Table 1A). Sensitivity, specificity, PPV, NPV and accuracy of GBS and AIMS65 for PLOS has been shown in Table 1B. Conclusion: AIMS65 is a simple scoring system, originally developed to predict mortality in patients with UGIB. This study provides preliminary evidence for the use of AIMS65 to predict longer LOS and ICU admission in patients with NVUGIB taking DAPT which may help resource allocation and prognostication. However, the study needs to be validated in a larger study population before it can be used in clinical practice.Table 1.: Comparison of Inpatient Outcomes of NVUGIB Patients.

Aims The shock index (SI), the modified shock index (MSI), and the age shock index (ASI) are simple scales which can be obtained bedside. These scales have been studied to predict outcomes in various diseases. We aimed to evaluate the use of SI, MSI and ASI in patients hospitalized with acute coronary syndrome (ACS). Method and Result In this cross-sectional study, 102 patients who were admitted with ACS to Kediri General Hospital between January and June 2020 were included. Patients who presented with shock were excluded. We calculated for each patient the admission SI (defined as heart rate/systolic blood pressure), MSI (heart rate/mean arterial pressure), and ASI (age × SI). The mean SI was 0.70 ± 0.20, MSI was 0.95 ± 0.23 and ASI was 42.33 ± 14.17. Receiver-operating characteristic (ROC) curve analysis was used to assess the ability of SI, MSI and ASI to predict in-hospital mortality and ICU admission. The observed AUC of SI was 0.706 (95% CI 0.572-0.840), MSI was 0.705 (95% CI 0.58-0.83), and ASI was 0.687 (95% CI 0.540-0.834) in predicting in-hospital mortality. The cut-off point for SI was 0.75 with sensitivity of 69.2% and specificity of 67.4% in predicting in-hospital mortality. MSI demonstrated better C-statistics (0.665, 95% CI = 0.565 - 0.765) in predicting ICU admission. Conclusion SI had the best predictive value for in-hospital mortality compared with MSI and ASI, while MSI could predict ICU admission in patients with ACS.

To identify factors predictive of admission of patients with cancer to an ICU. In addition, the study aimed to describe the characteristics and outcomes, both short-term and long-term, of patients with cancer admitted to the ICU. Retrospective case-control study, utilizing the institution's cancer registry. Comprehensive cancer center. Patients with cancer. The case group consisted of patients who required ICU admission during the study period, whereas the control group consisted of patients who did not require ICU admission. None. The patient characteristics and outcomes were recorded. Univariate and multivariate analyses were conducted to determine factors associated with ICU admission. The registry included 10,792 patients, and among those, 2,439 patients (22.6%) required ICU admission after a median of 10.1 months (interquartile range, 3.28-25.2). The following factors were associated with ICU admission: hematologic malignancy (odds ratio, 1.51; 95% CI, 1.26-1.81), chemotherapy (odds ratio, 1.74; 95% CI, 1.48-2.03), advanced cancer (odds ratio, 2.57; 95% CI, 1.44-4.60), and smoking (odds ratio, 1.38; 95% CI, 1.20-1.61). The most common ICU admission diagnoses were sepsis (21.5%) and respiratory insufficiency/failure (25.7%). The ICU mortality was 36.5%, whereas the 1-year and 5-year survival rates were 22.8% and 14.2%, respectively. In a comprehensive cancer center, about one fourth of the patients required ICU admission. Addressing modifiable risk factors associated with ICU admission is essential to potentially reduce ICU admissions and improve long-term survival.

Objective To identify factors predictive of admission of patients with cancer to an ICU. In addition, the study aimed to describe the characteristics and outcomes, both short-term and long-term, of patients with cancer admitted to the ICU. Design Retrospective case-control study, utilizing the institution's cancer registry. Setting Comprehensive cancer center. Patients Patients with cancer. The case group consisted of patients who required ICU admission during the study period, whereas the control group consisted of patients who did not require ICU admission. Intervention None. Measurements and main results The patient characteristics and outcomes were recorded. Univariate and multivariate analyses were conducted to determine factors associated with ICU admission. The registry included 10,792 patients, and among those, 2,439 patients (22.6%) required ICU admission after a median of 10.1 months (interquartile range, 3.28-25.2). The following factors were associated with ICU admission: hematologic malignancy (odds ratio, 1.51; 95% CI, 1.26-1.81), chemotherapy (odds ratio, 1.74; 95% CI, 1.48-2.03), advanced cancer (odds ratio, 2.57; 95% CI, 1.44-4.60), and smoking (odds ratio, 1.38; 95% CI, 1.20-1.61). The most common ICU admission diagnoses were sepsis (21.5%) and respiratory insufficiency/failure (25.7%). The ICU mortality was 36.5%, whereas the 1-year and 5-year survival rates were 22.8% and 14.2%, respectively. Conclusion In a comprehensive cancer center, about one fourth of the patients required ICU admission. Addressing modifiable risk factors associated with ICU admission is essential to potentially reduce ICU admissions and improve long-term survival.

<p><strong><em>Background</em></strong>: Scoring systems including CURB-65 and Pneumonia Severity Index (PSI) and novel or traditional biomarkers including procalcitonin (PCT) and c-reactive protein (CRP) are very significant for understanding the severity and prognosis in community-acquired pneumonia (CAP) patients, while prognostic items are useful for CAP prognostication and point-of-care decisions. The aim of this study was to investigate the usefulness of peripheral blood routine items in predicting ICU admission and 30-day mortality in CAP patients.</p> <p><strong><em>Methods</em></strong>: A retrospective study was conducted. All adult patients with a primary diagnosis of CAP were included and peripheral blood routine tests were evaluated. Univariate analysis and multivariate logistic regression analysis were used to explore association of risk factors with 30-day mortality among CAP patients. Receiver operating characteristic curves (ROC) were used to evaluate the sensitivity and specificity of peripheral blood routine items and compared with CURB-65 scores in predicting ICU admission and/or 30-day mortality.</p> <p><strong><em>Results</em></strong>: One hundred fifty patients were included and compared with non-ICU admission patients. There was a statistically significant difference in age, co-existing illness, RDW, WBC, and CURB-65 scores ranking in ICU admission patients (p < 0.05). In multivariate logistic regression analysis, we found RDW, WBC, and CURB-65 ≥ 3 scores increased the risk of 30-day mortality by 4.01, 1.65, and 3.43 times, respectively. The area under the curve (AUC) of ROC curves of RDW combined with WBC and CURB-65 was 0.786 (95% CI 0.701 to 0.876) and 0.836 (95% CI 0.764 to 0.908), respectively and the sensitivity was 84.0% and 60.0%, respectively, and the specificity 66.7% and 93.7%, respectively.</p> <p><strong><em>Conclusions</em></strong>: Elevated RDW and WBC increased mortality in adult CAP patients, RDW combined with WBC had a better sensitivity than CURB-65 scores in predicting ICU admission and/or mortality in CAP patients.</p>.

The CURB-65 scoring system is a simple tool in assessment and prognosis prediction for communityacquired pneumonia (CAP) patients. However, the variations in performance of CURB-65 in young and elderly patients, underestimation or overestimation of the severity, and mortality have often been reported. The aim of this study was to investigate the usefulness of serum high-sensitivity C reactive protein (hs-CRP) combined with CURB-65 in predicting ICU admission and 30-day mortality in CAP patients. We conducted a retrospective study. All patients over 18 years of age with a primary diagnosis of CAP were included, all of them received serum hs-CRP test and CURB-65 scaring evaluation. The main outcome measures were ICU admission and 30-day mortality. Receiver operating characteristic curves (ROC) were used to evaluate the sensitivity and specificity of the CURB-65 model and hs-CRP combined CURB-65 augmented model in predicting the main outcomes. Data from 150 patients was analyzed, in which the rate for patients requiring ICU admission was 30.67%, and the ultimate mortality rate was 24%. The areas of ROC curves (AUC) of CURB-65 was 0.859 (95% CI 0.705 to 1.000), hs-CRP combined CURB-65 augmented model was 0.864 (95% CI 0.692 to 1.000), ROC curve analyses showed the augmented model had higher sensitivity than the CURB-65 model in predicting main outcomes (p = 0.001). Measurement of serum hs-CRP in addition to the CURB-65 model improved the clinical usefulness in predicting ICU admission and mortality in CAP patients.

Background and aim Trauma-related thoracic injuries, particularly hemothorax and pulmonary contusion, frequently require critical care management. Timely identification of patients needing ICU admission is crucial for optimizing outcomes and resource use. This study aimed to identify clinical and radiological predictors of ICU admission in trauma patients presenting with hemothorax and/or pulmonary contusion. Patients and methods This retrospective cohort study included 200 trauma patients admitted to the Emergency Department of Esenyurt Necmi Kadioğlu State Hospital between January 1, 2023 and December 31, 2024. Demographic data, vital signs, laboratory parameters, imaging findings, and trauma scores were analyzed. Univariate and multivariate logistic regression analyses were used to determine independent predictors of ICU admission. Results Of the 200 patients, 88 (44%) required ICU admission. Hemoglobin levels were significantly lower in ICU-admitted patients and were found to be an independent predictor (P=0.023). The Glasgow Coma Scale showed a near-significant association (P=0.104). Trauma severity scores (ISS, RTS), vital signs, and trauma mechanism (blunt vs. penetrating) were not significantly associated with ICU need. Mortality was higher in the ICU group (17 vs. 6%). Conclusions Hemoglobin level is a significant predictor of ICU admission in trauma patients with hemothorax and/or pulmonary contusion. However, comprehensive clinical evaluation remains essential, as no single variable should determine ICU need. Prospective multicenter studies are recommended to validate these findings and support the development of reliable ICU admission criteria.

Theophylline remains the most widely pharmaceuticals for the treatment of acute and chronic asthma in several developing countries, as it is effective, cheap, and widely available. Few studies were investigated to predict the need of ICU admission based on clinical parameters recorded at admission. Hence, this study aimed to identify the predictors for ICU admission in acute theophylline intoxicated patients. It was carried out on one hundred and ten acutely theophylline poisoned patients who were admitted to Poison Control Unit, Emergency Hospital, Tanta University over a period of two years. For each patient, full sociodemographic, toxicological, clinical examination and routine laboratory investigations & serum theophylline level were done. Then, all findings of acute theophylline poisoned patients were analyzed against ICU admission. Statistical significant associations were found between ICU admission and gender, dose, CNS manifestations (agitations, hallucinations and tremors), hypotension, serum potassium and serum theophylline level. Logistic regression of clinically relevant variable showed that, patients who presented with hallucination, agitation, or hypotension had an increased likelihood of requiring admission to ICU and could correctly predicted 98.2% of cases. ROC curve analysis of serum theophylline accuracy revealed that, serum level ≥ 37.5 mg/L is a fair predictor for ICU admission. It could be concluded that, in acute theophylline intoxicated patients, hallucination, agitation and hypotension could be considered as good predictors for ICU admission. While, patients who had serum theophylline level ≥ 37.5 mg/L should be admitted in ICU as high risk patients.

Dear Editor, In this issue, we read with interest the article by Maignan and colleagues 1 regarding predictive factors for ICU admission in poisoned patients. Deliberate drug poisoning (DDP) is indeed a significant public health threat, particularly considering the current prescription drug overdose epidemic in the United States 2. To their credit, the authors acknowledge the limitations of their data set with regard to ‘toxicological anamnesis’ as well as the issue of generalizability to other regions that do not routinely practice direct ICU admission and bypass the ED. However, we feel some aspects of the methodology deserve some comments with regard to the interpretation of their final model, particularly on the use of Glasgow Coma Scale (GCS) as a predictor and ICU admission as the end-point in the study. The major limitation to the study is the use of ICU admission as the main outcome, rather than objective adverse end-points such as shock or cardiac arrest. The likely reason for this was the retrospective nature of the study, which likely precluded the ability to definitively capture adverse events such as shock or arrhythmias. Unfortunately, in doing so, the authors' model derived predictors of ICU admission, which are hospital dependent and practitioner dependent, may be significantly impacted by hospital census and do not necessarily signify that an adverse event has occurred. For example, patients in their study who were admitted to the ICU actually had similar median vital signs when comparing ED arrival and ICU arrival. Thus, the model essentially furthers a practice based on fear of certain toxic exposures, such as ‘cardiac drugs’, without using objective criteria (e.g. electrocardiographic factors) 3 to guide practice. In addition, the authors retrospectively applied the GCS to their patients and tested this as a predictor of ICU admission, which remained significant in the final model. This approach presents a variety of significant problems. Firstly, GCS is very difficult to calculate retrospectively and is often not charted even when calculated. The authors state that they did not adjust the model for missing values, which is a limitation to their statistical approach to logistic regression. In addition, GCS score for any individual patient changes dynamically across the time from ingestion; so given a population of patients with varying times to presentation, this appears to be quite a moving target. Most importantly, we are concerned that use of the GCS to guide triage in poisoning would lead to gross over-admission of patients to the ICU, particularly considering simple ethanol-intoxicated patients who present with coma (i.e. GCS < 8) but clearly do not require the precious resources afforded by the ICU. In summary, we applaud the authors' efforts to derive a risk stratification scheme for DDP triage, but would strongly caution against the use of their derived tool for individual patient triage, nor would we recommend the use of GCS to guide admission to the ICU. We look forward to future prospective work from the authors to improve upon their preliminary work.

Predicting Need for ICU in Community-Acquired Pneumonia

Comparison between CURB-65, PSI, and SIPF scores as predictors of ICU admission and mortality in community-acquired pneumonia