Lung Injury Prediction Score Is Useful in Predicting Acute Respiratory Distress Syndrome and Mortality in Surgical Critical Care Patients

Using a clinical decision support system to reduce excess driving pressure: the ALARM trial

BACKGROUND

Patients at need for ventilation often are at risk of acute respiratory distress syndrome (ARDS). Although lung-protective ventilation strategies, including low driving pressure settings, are well known to improve outcomes, clinical practice often diverges from these strategies. A clinical decision support (CDS) system can improve adherence to current guidelines; moreover, the potential of a CDS to enhance adherence can possibly be further increased by combination with a nudge type intervention.

METHODS

A prospective cohort trial was conducted in patients at risk of ARDS admitted to an intensive care unit (ICU). Patients were assigned to control or intervention by their date of admission: First, the control group was included without changing anything in clinical practice. Next, the CDS was activated showing an alert in the patient data management system if driving pressure exceeded recommended values; additionally, data on the performance of the wards were sent to the healthcare professionals as the nudge intervention. The main hypothesis was that this combined intervention would lead to a significant decrease in excess driving pressure.

RESULTS

The 472 included patients (230 in the control group and 242 in the intervention group) consisted of 33% females. The median age was 64 years; median Sequential Organ Failure Assessment score was 8. There was a significant reduction in excess driving pressure in the augmented ventilation modes (0.28 ± 0.67 mbar vs. 0.14 ± 0.45 mbar, p = 0.012) but not the controlled mode (0.37 ± 0.83 mbar vs. 0.32 ± 0.8 mbar, p = 0.53). However, there was no significant difference between groups in mechanical power, the number of ventilator-free days, or the percentage of patients showing progression to ARDS. Although there was no difference in progression to ARDS, 28-day mortality was higher in the intervention group. Notably, the mean overall driving pressure across both groups was low (12.02 mbar ± 2.77).

CONCLUSIONS

In a population at risk of ARDS, a combined intervention of a clinical decision support system and a nudge intervention was shown to reduce the excessive driving pressure above 15 mbar in augmented but not in controlled modes of ventilation.

Predictive Role of Lung Injury Prediction Score in the Development of Acute Respiratory Distress Syndrome in Korea

PURPOSE

Early recognition and therapeutic intervention are important in patients at high risk of acute respiratory distress syndrome (ARDS). The lung injury prediction score (LIPS) has been used to predict ARDS development; however, it was developed based on the previous definition of ARDS. We investigated the predictive role of LIPS in ARDS development according to its Berlin definition in the Korean population.

MATERIALS AND METHODS

This was a retrospective study that enrolled adult patients admitted to the intensive care unit (ICU) at a single university-affiliated hospital in Korea from September 1, 2018, to August 31, 2019. LIPS at the time of ICU admission and the development of ARDS were evaluated.

RESULTS

Of the 548 enrolled patients, 33 (6.0%) fulfilled the Berlin ARDS definition. The LIPS for non-ARDS and ARDS groups were 4.96±3.05 and 8.53±2.45, respectively (p<0.001); it was significantly associated with ARDS development (odds ratio 1.48, 95% confidence interval, 1.29-1.69; p<0.001). LIPS >6 predicted the development of ARDS with a sensitivity of 84.8% and a specificity of 67.2% [area under the curve (AUC)=0.82]. A modified LIPS model adjusted for age and severity at ICU admission predicted ICU mortality in patients with ARDS (AUC=0.80), but not in those without ARDS (AUC=0.54).

CONCLUSION

LIPS predicted the development of ARDS as diagnosed by the Berlin definition in the Korean population. LIPS provides useful information for managing patients with ARDS.

Supervised machine learning for the early prediction of acute respiratory distress syndrome (ARDS)

PURPOSE

Acute respiratory distress syndrome (ARDS) is a serious respiratory condition with high mortality and associated morbidity. The objective of this study is to develop and evaluate a novel application of gradient boosted tree models trained on patient health record data for the early prediction of ARDS.

MATERIALS AND METHODS

9919 patient encounters were retrospectively analyzed from the Medical Information Mart for Intensive Care III (MIMIC-III) data base. XGBoost gradient boosted tree models for early ARDS prediction were created using routinely collected clinical variables and numerical representations of radiology reports as inputs. XGBoost models were iteratively trained and validated using 10-fold cross validation.

RESULTS

On a hold-out test set, algorithm classifiers attained area under the receiver operating characteristic curve (AUROC) values of 0.905 when tested for the detection of ARDS at onset and 0.827, 0.810, and 0.790 for the prediction of ARDS at 12-, 24-, and 48-h windows prior to onset, respectively.

CONCLUSION

Supervised machine learning predictions may help predict patients with ARDS up to 48 h prior to onset.

Towards Reliable ARDS Clinical Decision Support: ARDS Patient Analytics with Free-text and Structured EMR Data

In this work, we utilize a combination of free-text and structured data to build Acute Respiratory Distress Syndrome(ARDS) prediction models and ARDS phenotype clusters. We derived 'Patient Context Vectors' representing patientspecific contextual ARDS risk factors, utilizing deep-learning techniques on ICD and free-text clinical notes data. The Patient Context Vectors were combined with structured data from the first 24 hours of admission, such as vital signs and lab results, to build an ARDS patient prediction model and an ARDS patient mortality prediction model achieving AUC of 90.16 and 81.01 respectively. The ability of Patient Context Vectors to summarize patients' medical history and current conditions is also demonstrated by the automatic clustering of ARDS patients into clinically meaningful phenotypes based on comorbidities, patient history, and presenting conditions. To our knowledge, this is the first study to successfully combine free-text and structured data, without any manual patient risk factor curation, to build real-time ARDS prediction models.

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

BACKGROUND

To 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.

METHODS

A 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.

RESULTS

All 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.

CONCLUSIONS

This 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.

Perioperative anaesthetic management of patients with or at risk of acute distress respiratory syndrome undergoing emergency surgery

Patients undergoing emergency surgery may present with the acute respiratory distress syndrome (ARDS) or develop this syndrome postoperatively. The incidence of ARDS in the postoperative period is relatively low, but the impact of ARDS on patient outcomes and healthcare costs is relevant Aakre et.al (Mayo Clin Proc 89:181-9, 2014).The development of ARDS as a postoperative pulmonary complication (PPC) is associated with prolonged hospitalisation, longer duration of mechanical ventilation, increased intensive care unit length of stay and high morbidity and mortality Ball et.al (Curr Opin Crit Care 22:379-85, 2016). In order to mitigate the risk of ARDS after surgery, the anaesthetic management and protective mechanical ventilation strategies play an important role. In particular, a careful integration of general anaesthesia with neuraxial or locoregional techniques might promote faster recovery and reduce opioid consumption. In addition, the use of low tidal volume, minimising plateau pressure and titrating a low-moderate PEEP level based on the patient's need can improve outcome and reduce intraoperative adverse events. Moreover, perioperative management of ARDS patients includes specific anaesthesia and ventilator settings, hemodynamic monitoring, moderately restrictive fluid administration and pain control.The aim of this review is to provide an overview and evidence- and opinion-based recommendations concerning the management of patients at risk of and with ARDS who undergo emergency surgical procedures.

Rationale and design of a prospective, multicentre, randomised, conventional treatment-controlled, parallel-group trial to evaluate the efficacy and safety of ulinastatin in preventing acute respiratory distress syndrome in high-risk patients

INTRODUCTION

Acute respiratory distress syndrome (ARDS) is challenging in the intensive care unit (ICU). Although pharmacotherapy for ARDS has gained increasing attention, most trials have yielded negative results. Patients with ARDS have usually been recruited as subjects; the inflammatory reaction has already expanded into a cascade at this point, and its severity is sufficient to damage the lung parenchyma. This raises the question of whether early treatment can prevent ARDS and the associated lung injury. We hypothesise that ARDS is preventable in high-risk patients by administration of ulinastatin as an anti-inflammatory drug before ARDS onset, and we are performing a study to test ulinastatin, a protease inhibitor, versus treatment-as-usual in a group of patients at increased risk for ARDS.

METHODS AND ANALYSIS

This report presents the protocol for a multicentre, randomised, conventional treatment-controlled, parallel group study to prevent the development of ARDS using ulinastatin in high-risk patients. The study population will comprise patients at risk of ARDS in the ICU (≥18 years of age and Lung Injury Prediction Score of >4); patients with confirmed ARDS and some other conditions (immunodeficiency, use of some drugs, etc.) will be excluded. The enrolled patients will be randomly allocated to an ulinastatin group (ulinastatin will be intravenously administered every 8 hours for a total of 600 000 U/day for five consecutive days) or control group. The efficacy of ulinastatin in preventing ARDS development will be evaluated by the incidence rate of ARDS as the primary outcome; the secondary outcomes include the severity of ARDS, clinical outcome, extrapulmonary organ function and adverse events incurred by ulinastatin. Based on the results of preliminary studies and presuming the incidence of ARDS will decrease by 9% in high-risk patients, 880 patients are needed to obtain statistical power of 80%.

ETHICS AND DISSEMINATION

This study has been approved by the Peking University Third Hospital Medical Science Research Ethics Committee. The findings will be published in peer-reviewed journals and presented at national and international conferences.

TRIAL REGISTRATION NUMBER

NCT03089957; Pre-results.

High-frequency power of heart rate variability can predict the outcome of thoracic surgical patients with acute respiratory distress syndrome on admission to the intensive care unit: a prospective, single-centric, case-controlled study

BACKGROUND

The morbidity and mortality of acute respiratory distress syndrome (ARDS) remains high, and the strategic focus of ARDS research has shifted toward identifying patients at high risk of mortality early in the course of illness. This study intended to identify the heart rate variability (HRV) measure that can predict the outcome of patients with ARDS on admission to the surgical intensive care unit (SICU).

METHODS

Patients who had lung or esophageal cancer surgery were included either in the ARDS group (n = 21) if they developed ARDS after surgery or in the control group (n = 11) if they did not. The ARDS patients were further stratified into survivors and non-survivors subgroups according to their outcomes. HRV measures of the patients were used for statistical analysis.

RESULTS

The mean RR interval (mRRI), high-frequency power (HFP) and product of low-/high-frequency power ratio tidal volume and tidal volume (LHR*VT) were significantly lower (p < 0.05), while the normalized HFP to VT ratio (nHFP/VT) was significantly higher in the ARDS patients (p = 0.011). The total power (TP), low-frequency power (LFP), HFP and HFP/VT were all significantly higher in the non-survived ARDS patients, whereas Richmond Agitation-Sedation Scale (RASS) was significantly lower in the non-survived ARDS patients. After adjustment for RASS, age and gender, firth logistic regression analysis identified the HFP, TP as the significant independent predictors of mortality for ARDS patients.

CONCLUSIONS

The vagal modulation of thoracic surgical patients with ARDS was enhanced as compared to that of non-ARDS patients, and the non-survived ARDS patients had higher vagal activity than those of survived ARDS patients. The vagal modulation-related parameters such as TP and HFP were independent predictors of mortality in patients with ARDS on admission to the SICU, and the HFP was found to be the best predictor of mortality for those ARDS patients. Increased vagal modulation might be an indicator for poor prognosis in critically ill patients following thoracic surgery.