Dataset Information

APACHE III Outcome Prediction in Patients Admitted to the Intensive Care Unit with Sepsis Associated Acute Lung Injury.

ABSTRACT: Acute Physiology and Chronic Health Evaluation (APACHE) III score has been widely used for prediction of clinical outcomes in mixed critically ill patients. However, it has not been validated in patients with sepsis-associated acute lung injury (ALI). The aim of the study was to explore the calibration and predictive value of APACHE III in patients with sepsis-associated ALI.The study was a secondary analysis of a prospective randomized controlled trial investigating the efficacy of rosuvastatin in sepsis-associated ALI (Statins for Acutely Injured Lungs from Sepsis, SAILS). The study population was sepsis-related ALI patients. The primary outcome of the current study was the same as in the original trial, 60-day in-hospital mortality, defined as death before hospital discharge, censored 60 days after enrollment. Discrimination of APACHE III was assessed by calculating the area under the receiver operating characteristic (ROC) curve (AUC) with its 95% CI. Hosmer-Lemeshow goodness-of-fit statistic was used to assess the calibration of APACHE III. The Brier score was reported to represent the overall performance of APACHE III in predicting outcome.A total of 745 patients were included in the study, including 540 survivors and 205 non-survivors. Non-survivors were significantly older than survivors (59.71 ± 16.17 vs 52.00 ± 15.92 years, p < 0.001). The primary causes of ALI were also different between survivors and non-survivors (p = 0.017). Survivors were more likely to have the cause of sepsis than non-survivors (21.2% vs. 15.1%). APACHE III score was higher in non-survivors than in survivors (106.72 ± 27.30 vs. 88.42 ± 26.86; p < 0.001). Discrimination of APACHE III to predict mortality in ALI patients was moderate with an AUC of 0.68 (95% confidence interval: 0.64-0.73).this study for the first time validated the discrimination of APACHE III in sepsis associated ALI patients. The result shows that APACHE III score has moderate predictive value for in-hospital mortality among adults with sepsis-associated acute lung injury.

SUBMITTER: Zhang Z

PROVIDER: S-EPMC4589281 | biostudies-literature | 2015

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

APACHE III Outcome Prediction in Patients Admitted to the Intensive Care Unit with Sepsis Associated Acute Lung Injury.

Zhang Zhongheng Z Chen Kun K Chen Lin L

PloS one 20150930 9

<h4>Background and objective</h4>Acute Physiology and Chronic Health Evaluation (APACHE) III score has been widely used for prediction of clinical outcomes in mixed critically ill patients. However, it has not been validated in patients with sepsis-associated acute lung injury (ALI). The aim of the study was to explore the calibration and predictive value of APACHE III in patients with sepsis-associated ALI.<h4>Method</h4>The study was a secondary analysis of a prospective randomized controlled ...[more]

PMID: 26422633

Similar Datasets

Project description:ObjectiveThis study aims to evaluate the accuracy of the Acute Physiology and Chronic Health Evaluation (APACHE) II score on different days in predicting the mortality of critically ill patients to identify the best time point for the APACHE II score.MethodsThe demographic and clinical data are retrieved from the Medical Information Mart for Intensive Care (MIMIC)-IV dataset. APACHE II scores on days 1, 2, 3, 5, 7, 14, and 28 of hospitalization are calculated, and their performance is evaluated using the area under the receiver operating characteristic (AUROC) analysis. The cut-off for defining the high risk of mortality is determined using Youden's index. The APACHE II score on day 3 is the best time point to predict hospital mortality of ICU patients. The Hosmer-Lemeshow goodness-of-fit test is then applied to evaluate the calibration of the day 3 APACHE II score.ResultsWe recruited 6,374 eligible subjects from the MIMIC-IV database. Day 3 is the optimal time point for obtaining the APACHE II score to predict the hospital mortality of patients. The best cut-off for day 3 APACHE II score is 17. When APACHE II score ≥17, the sensitivity for the non-survivors and survivors is 92.8 and 82.2%, respectively, and the positive predictive value (PPV) is 23.1%. When APACHE II socre <17, the specificity for non-survivors and survivors is 90.1 and 80.2%, respectively, and the negative predictive value (NPV) is 87.8%. When day-3 APACHE II is used to predict the hospital mortality, the AUROC is 0.743 (P <0.001). In the ≥17 group, the sensitivity of non-survivors and survivors is 92.2 and 81.3%, respectively, and the PPV is 30.3%. In the <17 group, the specificity of non-survivors and survivors is 100.0 and 80.2%, respectively, and the NPV is 81.6%. The Hosmer-Lemeshow test indicated day-3 APACHE II has a high predicting the hospital mortality (X 2 = 6.198, P = 0.625, consistency = 79.4%). However, the day-1 APACHE II has a poor calibration in predicting the hospital mortality rate (X 2 = 294.898, P <0.001).ConclusionDay-3 APACHE II score is an optimal biomarker to predict the outcomes of ICU patients; 17 is the best cut-off for defining patients at high risk of mortality.

Project description:BackgroundsThe in-hospital mortality in lung cancer patients admitted to intensive care unit (ICU) is extremely high. This study intended to adopt machine learning algorithm models to predict in-hospital mortality of critically ill lung cancer for providing relative information in clinical decision-making.MethodsData were extracted from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) for a training cohort and data extracted from the Medical Information Mart for eICU Collaborative Research Database (eICU-CRD) database for a validation cohort. Logistic regression, random forest, decision tree, light gradient boosting machine (LightGBM), eXtreme gradient boosting (XGBoost), and an ensemble (random forest+LightGBM+XGBoost) model were used for prediction of in-hospital mortality and important feature extraction. The AUC (area under receiver operating curve), accuracy, F1 score and recall were used to evaluate the predictive performance of each model. Shapley Additive exPlanations (SHAP) values were calculated to evaluate feature importance of each feature.ResultsOverall, there were 653 (24.8%) in-hospital mortality in the training cohort, and 523 (21.7%) in-hospital mortality in the validation cohort. Among the six machine learning models, the ensemble model achieved the best performance. The top 5 most influential features were the sequential organ failure assessment (SOFA) score, albumin, the oxford acute severity of illness score (OASIS) score, anion gap and bilirubin in random forest and XGBoost model. The SHAP summary plot was used to illustrate the positive or negative effects of the top 15 features attributed to the XGBoost model.ConclusionThe ensemble model performed best and might be applied to forecast in-hospital mortality of critically ill lung cancer patients, and the SOFA score was the most important feature in all models. These results might offer valuable and significant reference for ICU clinicians' decision-making in advance.

Project description:BACKGROUND:Bacterial infections are frequent triggers for diabetic ketoacidosis. In this context, delayed antibiotic treatment is associated with increased morbidity and mortality. Unnecessary administration of antimicrobial therapy might however, also negatively impact the prognosis. The usefulness of sepsis markers in diabetic ketoacidosis has not been assessed. Thus, we sought to investigate diagnostic performances of clinical and biological sepsis markers during diabetic ketoacidosis. METHODS:In this monocentric retrospective cohort study, all consecutive episodes of diabetic ketoacidosis (defined as pH ? 7.25, glycaemia > 300 mg/dL and presence of ketones) admitted in intensive care unit were included. A proven bacterial infection was defined as bacteriological documentation on any bacterial sample. Clinical (presence of fever: temperature > 38 °C and presence of hypothermia: temperature < 36 °C) and biological markers (whole blood count, neutrophils count, neutrophils-to-lymphocytes count ratio and procalcitonin), recorded at admission, were compared according to the presence or absence of a proven bacterial infection. RESULTS:Between 2011 and 2018, among 134 episodes of diabetic ketoacidosis, 102 were included (91 patients). Twenty out of 102 were infected. At admission, procalcitonin (median: 3.58 ng/mL vs 0.52 ng/mL, p?<?0.001) and presence of fever (25% vs 44%, p?=?0.007) were different between episodes with and without proven bacterial infection in both univariate and multivariate analysis. Whole blood count, neutrophils count, neutrophils-to-lymphocytes count ratio and presence of hypothermia were not different between both groups. The diagnostic performance analysis for procalcitonin revealed an area under the curve of 0.87 with an optimal cutoff of 1.44 ng/mL leading to a sensitivity of 0.90 and a specificity of 0.76. Combining procalcitonin and presence of fever allowed to distinguish proven bacterial infection episodes from those without proven bacterial infection. Indeed, all patients with procalcitonin level of more than 1.44 ng/mL and fever had proven bacterial infection episodes. The presence of one of these 2 markers was associated with 46% of proven bacterial infection episodes. No afebrile patient with procalcitonin level less than 1.44 ng/mL had a proven bacterial infection. CONCLUSION:At admission, combining procalcitonin and presence of fever may be of value to distinguish ketoacidosis patients with and without proven bacterial infection, admitted in intensive care unit.

Project description:BackgroundWe aimed to develop an early warning system for real-time sepsis prediction in the ICU by machine learning methods, with tools for interpretative analysis of the predictions. In particular, we focus on the deployment of the system in a target medical center with small historical samples.MethodsLight Gradient Boosting Machine (LightGBM) and multilayer perceptron (MLP) were trained on Medical Information Mart for Intensive Care (MIMIC-III) dataset and then finetuned on the private Historical Database of local Ruijin Hospital (HDRJH) using transfer learning technique. The Shapley Additive Explanations (SHAP) analysis was employed to characterize the feature importance in the prediction inference. Ultimately, the performance of the sepsis prediction system was further evaluated in the real-world study in the ICU of the target Ruijin Hospital.ResultsThe datasets comprised 6891 patients from MIMIC-III, 453 from HDRJH, and 67 from Ruijin real-world data. The area under the receiver operating characteristic curves (AUCs) for LightGBM and MLP models derived from MIMIC-III were 0.98 - 0.98 and 0.95 - 0.96 respectively on MIMIC-III dataset, and, in comparison, 0.82 - 0.86 and 0.84 - 0.87 respectively on HDRJH, from 1 to 5 h preceding. After transfer learning and ensemble learning, the AUCs of the final ensemble model were enhanced to 0.94 - 0.94 on HDRJH and to 0.86 - 0.9 in the real-world study in the ICU of the target Ruijin Hospital. In addition, the SHAP analysis illustrated the importance of age, antibiotics, net balance, and ventilation for sepsis prediction, making the model interpretable.ConclusionsOur machine learning model allows accurate real-time prediction of sepsis within 5-h preceding. Transfer learning can effectively improve the feasibility to deploy the prediction model in the target cohort, and ameliorate the model performance for external validation. SHAP analysis indicates that the role of antibiotic usage and fluid management needs further investigation. We argue that our system and methodology have the potential to improve ICU management by helping medical practitioners identify at-sepsis-risk patients and prepare for timely diagnosis and intervention.Trial registrationNCT05088850 (retrospectively registered).

Project description:BACKGROUND:This retrospective, multicenter study aimed to reveal risk predictors for mortality in the intensive care unit (ICU) as well as survival after ICU discharge in patients with acute myeloid leukemia (AML) requiring treatment in the ICU. METHODS AND RESULTS:Multivariate analysis of data for 187 adults with AML treated in the ICU in one institution revealed the following as independent prognostic factors for death in the ICU: arterial oxygen partial pressure below 72 mmHg, active AML and systemic inflammatory response syndrome upon ICU admission, and need for hemodialysis and mechanical ventilation in the ICU. Based on these variables, we developed an ICU mortality score and validated the score in an independent cohort of 264 patients treated in the ICU in three additional tertiary hospitals. Compared with the Simplified Acute Physiology Score (SAPS) II, the Logistic Organ Dysfunction (LOD) score, and the Sequential Organ Failure Assessment (SOFA) score, our score yielded a better prediction of ICU mortality in the receiver operator characteristics (ROC) analysis (AUC = 0.913 vs. AUC = 0.710 [SAPS II], AUC = 0.708 [LOD], and 0.770 [SOFA] in the training cohort; AUC = 0.841 for the developed score vs. AUC = 0.730 [SAPSII], AUC = 0.773 [LOD], and 0.783 [SOFA] in the validation cohort). Factors predicting decreased survival after ICU discharge were as follows: relapse or refractory disease, previous allogeneic stem cell transplantation, time between hospital admission and ICU admission, time spent in ICU, impaired diuresis, Glasgow Coma Scale <8 and hematocrit of ?25% at ICU admission. Based on these factors, an ICU survival score was created and used for risk stratification into three risk groups. This stratification discriminated distinct survival rates after ICU discharge. CONCLUSIONS:Our data emphasize that although individual risks differ widely depending on the patient and disease status, a substantial portion of critically ill patients with AML benefit from intensive care.

Dataset Information

APACHE III Outcome Prediction in Patients Admitted to the Intensive Care Unit with Sepsis Associated Acute Lung Injury.

Publications

APACHE III Outcome Prediction in Patients Admitted to the Intensive Care Unit with Sepsis Associated Acute Lung Injury.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets