Predictive model for acute respiratory distress syndrome events in ICU patients in China using machine learning algorithms: a secondary analysis of a cohort study

BackgroundTo develop a machine learning model for predicting acute respiratory distress syndrome (ARDS) events through commonly available parameters, including baseline characteristics and clinical and laboratory parameters.MethodsA secondary analysis of a multi-centre prospective observational cohort study from five hospitals in Beijing, China, was conducted from January 1, 2011, to August 31, 2014. A total of 296 patients at risk for developing ARDS admitted to medical intensive care units (ICUs) were included. We applied a random forest approach to identify the best set of predictors out of 42 variables measured on day 1 of admission.ResultsAll patients were randomly divided into training (80%) and testing (20%) sets. Additionally, these patients were followed daily and assessed according to the Berlin definition. The model obtained an average area under the receiver operating characteristic (ROC) curve (AUC) of 0.82 and yielded a predictive accuracy of 83%. For the first time, four new biomarkers were included in the model: decreased minimum haematocrit, glucose, and sodium and increased minimum white blood cell (WBC) count.ConclusionsThis newly established machine learning-based model shows good predictive ability in Chinese patients with ARDS. External validation studies are necessary to confirm the generalisability of our approach across populations and treatment practices.

[1]  Arthur S Slutsky,et al.  Acute Respiratory Distress Syndrome The Berlin Definition , 2012 .

[2]  G. Bernard,et al.  Lower tidal volume ventilation and plasma cytokine markers of inflammation in patients with acute lung injury* , 2005, Critical care medicine.

[3]  N. Harada,et al.  Rapid assay for plasma soluble E-selectin predicts the development of acute respiratory distress syndrome in patients with systemic inflammatory response syndrome. , 2006, Translational research : the journal of laboratory and clinical medicine.

[4]  M. Kenward,et al.  An Introduction to the Bootstrap , 2007 .

[5]  H. Tie,et al.  Risk of Acute Lung Injury/Acute Respiratory Distress Syndrome in Critically Ill Adult Patients with Pre-Existing Diabetes: A Meta-Analysis , 2014, PloS one.

[6]  L. Ware,et al.  Biomarkers in acute lung injury--marking forward progress. , 2011, Critical care clinics.

[7]  K. Tarte,et al.  Impaired efferocytosis and neutrophil extracellular trap clearance by macrophages in ARDS , 2018, European Respiratory Journal.

[8]  Jesús Blanco,et al.  Current incidence and outcome of the acute respiratory distress syndrome , 2016, Current opinion in critical care.

[9]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[10]  J. McNelis,et al.  Effect of hematocrit on regional oxygen delivery and extraction in an adult respiratory distress syndrome animal model. , 2000, American journal of surgery.

[11]  Daniel Brodie,et al.  Acute Respiratory Distress Syndrome: Advances in Diagnosis and Treatment , 2018, JAMA.

[12]  D. Slonim.  From patterns to pathways: gene expression data analysis comes of age , 2002, Nature Genetics.

[13]  Kevin Delucchi,et al.  Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials. , 2014, The Lancet. Respiratory medicine.

[14]  R. Hoover,et al.  Regulation of Lung Epithelial Sodium Channels by Cytokines and Chemokines , 2017, Front. Immunol..

[15]  Hilde van der Togt,et al.  Publisher's Note , 2003, J. Netw. Comput. Appl..

[16]  V. Herasevich,et al.  Acute lung injury prediction score: derivation and validation in a population-based sample , 2010, European Respiratory Journal.

[17]  G. Martin,et al.  Plasma Neutrophil Elastase and Elafin as Prognostic Biomarker for Acute Respiratory Distress Syndrome: A Multicenter Survival and Longitudinal Prospective Observation Study , 2017, Shock.

[18]  P. Pronovost,et al.  Short-term mortality prediction for acute lung injury patients: External validation of the Acute Respiratory Distress Syndrome Network prediction model* , 2011, Critical care medicine.

[19]  F. Cunha,et al.  α1-Acid Glycoprotein Decreases Neutrophil Migration and Increases Susceptibility to Sepsis in Diabetic Mice , 2012, Diabetes.

[20]  J. Tomashefski.  Pulmonary pathology of acute respiratory distress syndrome. , 2000, Clinics in chest medicine.

[21]  A. Dubin,et al.  Impact of positive end-expiratory pressure on the definition of acute respiratory distress syndrome , 2003, Intensive Care Medicine.

[22]  J. Sznajder,et al.  Protein kinase A-Iα regulates Na,K-ATPase endocytosis in alveolar epithelial cells exposed to high CO(2) concentrations. , 2013, American journal of respiratory cell and molecular biology.

[23]  Fernando Frutos-Vivar,et al.  Comparison of Clinical Criteria for the Acute Respiratory Distress Syndrome with Autopsy Findings , 2004, Annals of Internal Medicine.

[24]  M. Matthay,et al.  Therapeutic strategies for severe acute lung injury , 2010, Critical care medicine.

[25]  H. Wong,et al.  Adaptation of a Biomarker-Based Sepsis Mortality Risk Stratification Tool for Pediatric Acute Respiratory Distress Syndrome* , 2018, Critical care medicine.

[26]  G. Bernard,et al.  External validation of a biomarker and clinical prediction model for hospital mortality in acute respiratory distress syndrome , 2017, Intensive Care Medicine.

[27]  L. Liang,et al.  Whole blood microRNA markers are associated with acute respiratory distress syndrome , 2017, Intensive Care Medicine Experimental.

[28]  K. Wernecke,et al.  Nucleated red blood cells as predictors of mortality in patients with acute respiratory distress syndrome (ARDS): an observational study , 2018, Annals of Intensive Care.

[29]  Diane P. Martin,et al.  Incidence and outcomes of acute lung injury. , 2005, The New England journal of medicine.

[30]  G. Bernard,et al.  Biomarkers of lung epithelial injury and inflammation distinguish severe sepsis patients with acute respiratory distress syndrome , 2013, Critical Care.

[31]  P. Parsons,et al.  Use of risk reclassification with multiple biomarkers improves mortality prediction in acute lung injury , 2011, Critical care medicine.

[32]  L. Breiman.  Arcing Classifiers , 1998 .

[33]  J. Sznajder,et al.  Regulation of Na,K-ATPase during acute lung injury , 2007, Journal of bioenergetics and biomembranes.

[34]  M. Bonten,et al.  Plasma suPAR as a prognostic biological marker for ICU mortality in ARDS patients , 2015, Intensive Care Medicine.

[35]  M. A. Romera,et al.  A Risk Tertiles Model for Predicting Mortality in Patients With Acute Respiratory Distress Syndrome: Age, Plateau Pressure, and PaO2/FIO2 at ARDS Onset Can Predict Mortality , 2011, Respiratory Care.

[36]  S. T. Buckland,et al.  An Introduction to the Bootstrap. , 1994 .

[37]  G. Amerongen,et al.  Plasma Biomarkers for Acute Respiratory Distress Syndrome: A Systematic Review and Meta-Analysis* , 2014, Critical care medicine.

[38]  M. Kollef,et al.  The acute respiratory distress syndrome. , 1995, The New England journal of medicine.

[39]  P. Factor,et al.  Overexpression of the Na-K-ATPase alpha2-subunit improves lung liquid clearance during ventilation-induced lung injury. , 2008, American journal of physiology. Lung cellular and molecular physiology.

[40]  M. Matthay,et al.  Biomarkers in acute lung injury: insights into the pathogenesis of acute lung injury. , 2011, Critical care clinics.

[41]  T L Petty,et al.  Acute respiratory distress in adults. , 1967, Lancet.

[42]  Ó. Peñuelas,et al.  Comparison of the Berlin definition for acute respiratory distress syndrome with autopsy. , 2013, American journal of respiratory and critical care medicine.

[43]  J. Sznajder,et al.  Insulin regulates alveolar epithelial function by inducing Na+/K+-ATPase translocation to the plasma membrane in a process mediated by the action of Akt , 2010, Journal of Cell Science.

[44]  D. DeMets,et al.  Biomarkers and surrogate endpoints: Preferred definitions and conceptual framework , 2001, Clinical pharmacology and therapeutics.

[45]  Fatin Zaklouta,et al.  Traffic sign classification using K-d trees and Random Forests , 2011, The 2011 International Joint Conference on Neural Networks.

[46]  G. Downey,et al.  The fibroproliferative response in acute respiratory distress syndrome: mechanisms and clinical significance , 2013, European Respiratory Journal.

[47]  Joel L. Davis,et al.  Adaptive Critics and the Basal Ganglia , 1995 .

[48]  H. Ni,et al.  Prediction Model for Critically Ill Patients with Acute Respiratory Distress Syndrome , 2015, PloS one.

[49]  Lesly A. Dossett,et al.  Acute lung injury in patients with traumatic injuries: utility of a panel of biomarkers for diagnosis and pathogenesis. , 2010, The Journal of trauma.

[50]  Z. Bauman,et al.  Lung Injury Prediction Score Is Useful in Predicting Acute Respiratory Distress Syndrome and Mortality in Surgical Critical Care Patients , 2015, Critical care research and practice.

[51]  Lipo Wang.  Support vector machines : theory and applications , 2005 .

[52]  K. Famous,et al.  Acute Respiratory Distress Syndrome Subphenotypes Respond Differently to Randomized Fluid Management Strategy , 2016, American journal of respiratory and critical care medicine.

[53]  J. Sznajder,et al.  Hypoxia Leads to Na,K-ATPase Downregulation via Ca2+ Release-Activated Ca2+ Channels and AMPK Activation , 2011, Molecular and Cellular Biology.

[54]  O. Gajic,et al.  Predictive value of plasma biomarkers for mortality and organ failure development in patients with acute respiratory distress syndrome. , 2015, Journal of critical care.

[55]  Arthur S Slutsky,et al.  Biomarkers for the acute respiratory distress syndrome: how to make the diagnosis more precise. , 2017, Annals of translational medicine.

[56]  Ognjen Gajic,et al.  Early identification of patients at risk of acute lung injury: evaluation of lung injury prediction score in a multicenter cohort study. , 2011, American journal of respiratory and critical care medicine.

[57]  D. Christiani,et al.  A genome-wide expression analysis in blood identifies pre-elafin as a biomarker in ARDS. , 2008, American journal of respiratory cell and molecular biology.

[58]  Ruilan Wang,et al.  Ventilator management for acute respiratory distress syndrome associated with avian influenza A (H7N9) virus infection: A case series. , 2018, World journal of emergency medicine.

[59]  S. Jancar,et al.  Sepsis-Induced Acute Lung Injury (ALI) Is Milder in Diabetic Rats and Correlates with Impaired NFkB Activation , 2012, PloS one.