I-TRACH: Validating A Tool for Predicting Prolonged Mechanical Ventilation

Predicting Tracheostomy Need on Admission to the Intensive Care Unit-A Multicenter Machine Learning Analysis

OBJECTIVE

It is difficult to predict which mechanically ventilated patients will ultimately require a tracheostomy which further predisposes them to unnecessary spontaneous breathing trials, additional time on the ventilator, increased costs, and further ventilation-related complications such as subglottic stenosis. In this study, we aimed to develop a machine learning tool to predict which patients need a tracheostomy at the onset of admission to the intensive care unit (ICU).

STUDY DESIGN

Retrospective Cohort Study.

SETTING

Multicenter Study of 335 Intensive Care Units between 2014 and 2015.

METHODS

The eICU Collaborative Research Database (eICU-CRD) was utilized to obtain the patient cohort. Inclusion criteria included: (1) Age >18 years and (2) ICU admission requiring mechanical ventilation. The primary outcome of interest included tracheostomy assessed via a binary classification model. Models included logistic regression (LR), random forest (RF), and Extreme Gradient Boosting (XGBoost).

RESULTS

Of 38,508 invasively mechanically ventilated patients, 1605 patients underwent a tracheostomy. The XGBoost, RF, and LR models had fair performances at an AUROC 0.794, 0.780, and 0.775 respectively. Limiting the XGBoost model to 20 features out of 331, a minimal reduction in performance was observed with an AUROC of 0.778. Using Shapley Additive Explanations, the top features were an admission diagnosis of pneumonia or sepsis and comorbidity of chronic respiratory failure.

CONCLUSIONS

Our machine learning model accurately predicts the probability that a patient will eventually require a tracheostomy upon ICU admission, and upon prospective validation, we have the potential to institute earlier interventions and reduce the complications of prolonged ventilation.

Position Statement of ISCCM Committee on Weaning from Mechanical Ventilator

Background and purpose

Weaning from a mechanical ventilator is a milestone in the recovery of seriously ill patients in Intensive care. Failure to wean and re-intubation adversely affects the outcome. The method of mechanical ventilation (MV) varies between different ICUs and so does the practice of weaning. Therefore, updated guidelines based on contemporary literature are designed to guide intensivists in modern ICUs. This is the first ISCCM Consensus Statement on weaning complied by a committee on weaning. The recommendations are intended to be used by all the members of the ICU (Intensivists, Registrars, Nurses, and Respiratory Therapists).

Methods

A Committee on weaning from MV, formed by the Indian Society of Critical Care Medicine (ISCCM) has formulated this statement on weaning from mechanical ventilators in intensive care units (ICUs) after a review of the literature. Literature was first circulated among expert committee members and allotted sections to each member. Sections of the statement written by sectional authors were peer-reviewed on multiple occasions through virtual meetings. After the final manuscript is accepted by all the committee members, it is submitted for peer review by central guideline committee of ISCCM. Once approved it has passed through review by the Editorial Board of IJCCM before it is published here as "ISCCM consensus statement on weaning from mechanical ventilator". As per the standard accepted for all its guidelines of ISCCM, we followed the modified grading of recommendations assessment, development and evaluation (GRADE) system to classify the quality of evidence and strength of recommendation. Cost-benefit, risk-benefit analysis, and feasibility of implementation in Indian ICUs are considered by the committee along with the strength of evidence. Type of ventilators and their modes, ICU staffing pattern, availability of critical care nurses, Respiratory therapists, and day vs night time staffing are aspects considered while recommending for or against any aspect of weaning.

Result

This document makes recommendation on various aspects of weaning, namely, definition, timing, weaning criteria, method of weaning, diagnosis of failure to wean, defining difficult to wean, Use of NIV, HFOV as adjunct to weaning, role of tracheostomy in weaning, weaning in of long term ventilated patients, role of physiotherapy, mobilization in weaning, Role of nutrition in weaning, role of diaphragmatic ultrasound in weaning prediction etc. Out of 42 questions addressed; the committee provided 39 recommendations and refrained from 3 questions. Of these 39; 32 are based on evidence and 7 are based on expert opinion of the committee members. It provides 27 strong recommendations and 12 weak recommendations (suggestions).

Conclusion

This guideline gives extensive review on weaning from mechanical ventilator and provides various recommendations on weaning from mechanical ventilator. Though all efforts are made to make is as updated as possible one needs to review any guideline periodically to keep it in line with upcoming concepts and standards.

How to cite this article

Clerk AM, Shah RJ, Kothari J, Sodhi K, Vadi S, Bhattacharya PK, et al. Position Statement of ISCCM Committee on Weaning from Mechanical Ventilator. Indian J Crit Care Med 2024;28(S2):S233-S248.

The impact of comorbidities on prolonged mechanical ventilation in patients with chronic obstructive pulmonary disease

BACKGROUND

In patients with chronic obstructive pulmonary disease (COPD) and acute respiratory failure, approximately 10% of them are considered to be at high risk for prolonged mechanical ventilation (PMV, > 21 days). PMV have been identified as independent predictors of unfavorable outcomes. Our previous study revealed that patients aged 70 years older and COPD severity were at a significantly higher risk for PMV. We aimed to analyze the impact of comorbidities and their associated risks in patients with COPD who require PMV.

METHODS

The data used in this study was collected from Kaohsiung Medical University Hospital Research Database. The COPD subjects were the patients first diagnosed COPD (index date) between January 1, 2012 and December 31, 2020. The exclusion criteria were the patients with age less than 40 years, PMV before the index date or incomplete records. COPD and non-COPD patients, matched controls were used by applying the propensity score matching method.

RESULTS

There are 3,744 eligible patients with COPD in the study group. The study group had a rate of 1.6% (60 cases) patients with PMV. The adjusted HR of PMV was 2.21 (95% CI 1.44-3.40; P < 0.001) in the COPD patients than in non-COPD patients. Increased risks of PMV were found significantly for patients with diabetes mellitus (aHR 4.66; P < 0.001), hypertension (aHR 3.20; P = 0.004), dyslipidemia (aHR 3.02; P = 0.015), congestive heart failure (aHR 6.44; P < 0.001), coronary artery disease (aHR 3.11; P = 0.014), stroke (aHR 6.37; P < 0.001), chronic kidney disease (aHR 5.81 P < 0.001) and Dementia (aHR 5.78; P < 0.001).

CONCLUSIONS

Age, gender, and comorbidities were identified as significantly higher risk factors for PMV occurrence in the COPD patients compared to the non-COPD patients. Beyond age, comorbidities also play a crucial role in PMV in COPD.

Early risk factors for prolonged mechanical ventilation in patients with severe blunt thoracic trauma: A retrospective cohort study

BACKGROUND

A significant proportion of patients with severe chest trauma require mechanical ventilation (MV). Early prediction of the duration of MV may influence clinical decisions. We aimed to determine early risk factors for prolonged MV among adults suffering from severe blunt thoracic trauma.

METHODS

This retrospective, single-center, cohort study included all patients admitted between January 2014 and December 2020 due to severe blunt chest trauma. The primary outcome was prolonged MV, defined as invasive MV lasting more than 14 days. Multivariable logistic regression was performed to identify independent risk factors for prolonged MV.

RESULTS

The final analysis included 378 patients. The median duration of MV was 9.7 (IQR 3.0-18.0) days. 221 (58.5 %) patients required MV for more than 7 days and 143 (37.8 %) for more than 14 days. Male gender (aOR 3.01, 95 % CI 1.63-5.58, p < 0.001), age (aOR 1.40, 95 % CI 1.21-1.63, p < 0.001, for each category above 30 years), presence of severe head trauma (aOR 3.77, 95 % CI 2.23-6.38, p < 0.001), and transfusion of >5 blood units on admission (aOR 2.85, 95 % CI 1.62-5.02, p < 0.001) were independently associated with prolonged MV. The number of fractured ribs and the extent of lung contusions were associated with MV for more than 7 days, but not for 14 days. In the subgroup of 134 patients without concomitant head trauma, age (aOR 1.63, 95 % CI 1.18-2.27, p = 0.004, for each category above 30 years), respiratory comorbidities (aOR 9.70, 95 % CI 1.49-63.01, p = 0.017), worse p/f ratio during the first 24 h (aOR 1.55, 95 % CI 1.15-2.09, p = 0.004), and transfusion of >5 blood units on admission (aOR 5.71 95 % CI 1.84-17.68, p = 0.003) were independently associated with MV for more than 14 days.

CONCLUSIONS

Several predictors have been identified as independently associated with prolonged MV. Patients who meet these criteria are at high risk for prolonged MV and should be considered for interventions that could potentially shorten MV duration and reduce associated complications. Hemodynamically stable, healthy young patients suffering from severe thoracic trauma but no head injury, including those with extensive lung contusions and rib fractures, have a low risk of prolonged MV.

Machine Learning Algorithms Predict Successful Weaning From Mechanical Ventilation Before Intubation: Retrospective Analysis From the Medical Information Mart for Intensive Care IV Database

BACKGROUND

The prediction of successful weaning from mechanical ventilation (MV) in advance of intubation can facilitate discussions regarding end-of-life care before unnecessary intubation.

OBJECTIVE

We aimed to develop a machine learning-based model that predicts successful weaning from ventilator support based on routine clinical and laboratory data taken before or immediately after intubation.

METHODS

We used the Medical Information Mart for Intensive Care IV database, which is an open-access database covering 524,740 admissions of 382,278 patients in Beth Israel Deaconess Medical Center, United States, from 2008 to 2019. We selected adult patients who underwent MV in the intensive care unit (ICU). Clinical and laboratory variables that are considered relevant to the prognosis of the patient in the ICU were selected. Data collected before or within 24 hours of intubation were used to develop machine learning models that predict the probability of successful weaning within 14 days of ventilator support. Developed models were integrated into an ensemble model. Performance metrics were calculated by 5-fold cross-validation for each model, and a permutation feature importance and Shapley additive explanations analysis was conducted to better understand the impacts of individual variables on outcome prediction.

RESULTS

Of the 23,242 patients, 19,025 (81.9%) patients were successfully weaned from MV within 14 days. Using the preselected 46 clinical and laboratory variables, the area under the receiver operating characteristic curve of CatBoost classifier, random forest classifier, and regularized logistic regression classifier models were 0.860 (95% CI 0.852-0.868), 0.855 (95% CI 0.848-0.863), and 0.823 (95% CI 0.813-0.832), respectively. Using the ensemble voting classifier using the 3 models above, the final model revealed the area under the receiver operating characteristic curve of 0.861 (95% CI 0.853-0.869), which was significantly better than that of Simplified Acute Physiology Score II (0.749, 95% CI 0.742-0.756) and Sequential Organ Failure Assessment (0.588, 95% CI 0.566-0.609). The top features included lactate and anion gap. The model's performance achieved a plateau with approximately the top 21 variables.

CONCLUSIONS

We developed machine learning algorithms that can predict successful weaning from MV in advance to intubation in the ICU. Our models can aid the appropriate management for patients who hesitate to decide on ventilator support or meaningless end-of-life care.

Prediction prolonged mechanical ventilation in trauma patients of the intensive care unit according to initial medical factors: a machine learning approach

The goal of this study was to develop a predictive machine learning model to predict the risk of prolonged mechanical ventilation (PMV) in patients admitted to the intensive care unit (ICU), with a focus on laboratory and Arterial Blood Gas (ABG) data. This retrospective cohort study included ICU patients admitted to Rajaei Hospital in Shiraz between 2016 and March 20, 2022. All adult patients requiring mechanical ventilation and seeking ICU admission had their data analyzed. Six models were created in this study using five machine learning models (PMV more than 3, 5, 7, 10, 14, and 23 days). Patients' demographic characteristics, Apache II, laboratory information, ABG, and comorbidity were predictors. This study used Logistic regression (LR), artificial neural networks (ANN), support vector machines (SVM), random forest (RF), and C.5 decision tree (C.5 DT) to predict PMV. The study enrolled 1138 eligible patients, excluding brain-dead patients and those without mechanical ventilation or a tracheostomy. The model PMV > 14 days showed the best performance (Accuracy: 83.63-98.54). The essential ABG variables in our two optimal models (artificial neural network and decision tree) in the PMV > 14 models include FiO2, paCO2, and paO2. This study provides evidence that machine learning methods outperform traditional methods and offer a perspective for achieving a consensus definition of PMV. It also introduces ABG and laboratory information as the two most important variables for predicting PMV. Therefore, there is significant value in deploying such models in clinical practice and making them accessible to clinicians to support their decision-making.

Developing an explainable machine learning model to predict the mechanical ventilation duration of patients with ARDS in intensive care units

BACKGROUND

Acute respiratory distress syndrome (ARDS) is common in intensive care units with high mortality rate and mechanical ventilation (MV) is the most important related treatment. Early prediction of MV duration has benefit for patients risk stratification and care strategies support.

OBJECTIVE

To develop an explainable model for predicting mechanical ventilation (MV) duration in patients with ARDS using the machine learning (ML) approach.

METHOD

The number of 1,148, 1,697, and 29 ARDS patients admitted to intensive care units (ICU) in the MIMIC-IV, eICU-CRD, and AmsterdamUMCdb databases were included in the study. Features at MV initiation from the MIMIC-IV dataset were used to train prediction models based on seven supervised machine learning algorithms. After 5-fold cross-validation for hyperparameters tuning, the hyperparameters- optimized model of different algorithms was tested by external datasets extracted from eICU-CRD and Amsterdamumcdb. Finally, three descriptive machine learning explanation methods were conducted for the model explanation.

RESULT

The XGBoosting model showed the most stable and accurate performance among two testing datasets (RMSE= 5.57 and 5.46 days in eICU-CRD and AmsterdamUMCdb) and was selected as the optimal model. The model explanation based on SHAP, LIME, and DALEX results showed a consistent result, vasopressor, PH, and SOFA score had the highest effect on MV duration prediction.

CONCLUSION

ML models with features at MV initiation can accurate predict MV duration in patients with ARDS in ICUs. Among seven algorithms, XGB models showed the best performance (RMSE= 5.57 and 5.46 in two external datasets). LIME, SHAP, and Breakdown methods showed good performance as AXI methods.

Alveolar epithelial glycocalyx degradation mediates surfactant dysfunction and contributes to acute respiratory distress syndrome

Acute respiratory distress syndrome (ARDS) is a common cause of respiratory failure yet has few pharmacologic therapies, reflecting the mechanistic heterogeneity of lung injury. We hypothesized that damage to the alveolar epithelial glycocalyx, a layer of glycosaminoglycans interposed between the epithelium and surfactant, contributes to lung injury in patients with ARDS. Using mass spectrometry of airspace fluid noninvasively collected from mechanically ventilated patients, we found that airspace glycosaminoglycan shedding (an index of glycocalyx degradation) occurred predominantly in patients with direct lung injury and was associated with duration of mechanical ventilation. Male patients had increased shedding, which correlated with airspace concentrations of matrix metalloproteinases. Selective epithelial glycocalyx degradation in mice was sufficient to induce surfactant dysfunction, a key characteristic of ARDS, leading to microatelectasis and decreased lung compliance. Rapid colorimetric quantification of airspace glycosaminoglycans was feasible and could provide point-of-care prognostic information to clinicians and/or be used for predictive enrichment in clinical trials.

Factors associated with prolonged weaning from mechanical ventilation in medical patients

BACKGROUND

Patients who need prolonged mechanical ventilation (MV) have high resource utilization and relatively poor outcomes. The pathophysiologic mechanisms leading to weaning failure in this group may be complex and multifactorial. The aim of this study was to investigate the factors associated with prolonged weaning based on the Weaning Outcome according to a New Definition (WIND) classification.

METHODS

This is a prospective observational study with consecutive adult patients receiving MV for at least two calendar days in medical intensive care units from 1 November 2017 to 30 September 2020. Eligible patients were divided in a non-prolonged weaning group, including short and difficult weaning, and in a prolonged weaning group according to the WIND classification. The risk factors at the time of first separation attempt associated with prolonged weaning were analyzed using a multivariable logistic regression model.

RESULTS

Of the total 915 eligible patients, 172 (18.8%) patients were classified as prolonged weaning. A higher proportion of the prolonged weaning group had previous histories of endotracheal intubation, chronic lung disease, and hematologic malignancies. When compared with the non-prolonged weaning group, the median duration of MV before the first spontaneous breathing trial (SBT) was longer and the proportion of tracheostomized patients was higher in prolonged weaning group. In addition, the prolonged weaning group used higher peak inspiratory pressures and yielded lower PaO₂/FiO₂ ratios at the day of the first SBT compared with the non-prolonged weaning group. In multivariate analyses, the duration of MV before first SBT (adjusted odds ratio [OR] = 1.14, 95% confidence interval [CI] = 1.06-1.22, p < 0.001), tracheostomy state (adjusted OR = 1.95, 95% CI = 1.04-3.63, p = 0.036), PaO₂/FiO₂ ratio (adjusted OR = 1.00, 95% CI = 0.99-1.00, p = 0.023), and need for renal replacement therapy (adjusted OR = 2.68, 95% CI = 1.16-6.19, p = 0.021) were independently associated with prolonged weaning. After the exclusion of patients who underwent tracheostomy before the SBTs, similar results were obtained.

CONCLUSION

Longer duration of MV before the first SBT, tracheostomy status, poor oxygenation, and need for renal replacement therapy at the time of first SBT can predict prolonged weaning.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier NCT05134467.

Predictors of prolonged mechanical ventilation identified at an emergency visit for elderly people: A retrospective cohort study

The aim of this study was to determine the factors that are associated with prolonged mechanical ventilation in elderly patients.Retrospective cohort studySingle tertiary hospital in JapanWe retrospectively identified 228 patients aged 75 years or older who were admitted to a single tertiary care center in Japan between January 1, 2014 and December 31, 2017 because of endogenous diseases and underwent mechanical ventilation.The primary outcome was extubation difficulty, which was defined as the need for mechanical ventilation for more than 14 days after intubation, reintubation within 72 hours after extubation, tracheotomy or extubation, or death within 14 days after intubation.A multivariate analysis showed that age (odds ratio [OR] = 0.95; 95% confidence interval [CI] = 0.66-1.38; P = .80), gender (OR = 0.56; 95%CI = 0.27-1.17; P = .13), body mass index (BMI) (OR = 1.05; 95%CI = 0.98-1.14; P = .16), smoking history (OR = 0.64; 95%CI = 0.29-1.41; P = .27), Activities of daily living (ADL) (OR = 0.95; 95%CI = 0.49-1.83; P = .87), and modified acute physiology and chronic health evaluation (APACHE) II score (OR = 1.02; 95%CI = 0.95-1.09; P = .61) were not statistically significantly different. However, there were statistically significant differences in extubation difficulty between patients with diabetes mellitus (OR = 2.3; 95%CI = 1.01-5.12; P = .04) and those with cardiovascular disease diagnosis on admission (OR = 0.31; 95%CI = 0.1-0.97; P = .04).Diabetes mellitus and cardiovascular disease diagnosis on admission were factors that were associated with prolonged mechanical ventilation in the elderly. The results of this study may help to support shared decision making with patients or surrogate decision makers at the start of intensive care in the elderly.

Obesity as a Predictor of Prolonged Mechanical Ventilation

Objective

To examine the effect of including obesity with parameters of the I-TRACH scale in predicting the need for prolonged mechanical ventilation.

Study Design

A retrospective cohort study.

Setting

Tertiary care academic medical center.

Subjects and Methods

Consecutive patients were identified retrospectively over a 45-month period based on need for mechanical ventilation in the medical intensive care unit. Chart review was performed to collect demographic information as well as clinical data, including duration of mechanical ventilation, body mass index (BMI), and I-TRACH parameters (heart rate >110, serum urea nitrogen >25, serum pH <7.25, serum creatinine >2, serum bicarbonate <20). Statistical analysis was performed to identify any predictors of prolonged mechanical ventilation, defined as ≥14 days and as ≥10 days.

Results

In total, 455 patients were identified, with an average duration of mechanical ventilation of 10.4 days (range, 0-248 days). On univariate and multivariate regression analysis, only BMI >30 reached statistical significance with respect to prolonged mechanical ventilation ( P < .05). The I-TRACH parameters—either alone or in combination—were not significantly predictive.

Conclusion

This study challenges previous findings regarding the I-TRACH scale and the relation of its parameters to prolonged mechanical ventilation. Furthermore, BMI >30 alone was predictive of prolonged intubation. Inclusion of BMI in predictive models could assist current decision making in determining the likelihood of prolonged mechanical ventilation in medical intensive care unit patients going forward, and obesity should be considered a predictor of prolonged mechanical ventilation.