Predicting responders to prone positioning in mechanically ventilated patients with COVID-19 using machine learning.

BACKGROUND: For mechanically ventilated critically ill COVID-19 patients, prone positioning has quickly become an important treatment strategy, however, prone positioning is labor intensive and comes with potential adverse effects. Therefore, identifying which critically ill intubated COVID-19 patients will benefit may help allocate labor resources. METHODS: From the multi-center Dutch Data Warehouse of COVID-19 ICU patients from 25 hospitals, we selected all 3619 episodes of prone positioning in 1142 invasively mechanically ventilated patients. We excluded episodes longer than 24 h. Berlin ARDS criteria were not formally documented. We used supervised machine learning algorithms Logistic Regression, Random Forest, Naive Bayes, K-Nearest Neighbors, Support Vector Machine and Extreme Gradient Boosting on readily available and clinically relevant features to predict success of prone positioning after 4 h (window of 1 to 7 h) based on various possible outcomes. These outcomes were defined as improvements of at least 10% in PaO2/FiO2 ratio, ventilatory ratio, respiratory system compliance, or mechanical power. Separate models were created for each of these outcomes. Re-supination within 4 h after pronation was labeled as failure. We also developed models using a 20 mmHg improvement cut-off for PaO2/FiO2 ratio and using a combined outcome parameter. For all models, we evaluated feature importance expressed as contribution to predictive performance based on their relative ranking. RESULTS: The median duration of prone episodes was 17 h (11-20, median and IQR, N = 2632). Despite extensive modeling using a plethora of machine learning techniques and a large number of potentially clinically relevant features, discrimination between responders and non-responders remained poor with an area under the receiver operator characteristic curve of 0.62 for PaO2/FiO2 ratio using Logistic Regression, Random Forest and XGBoost. Feature importance was inconsistent between models for different outcomes. Notably, not even being a previous responder to prone positioning, or PEEP-levels before prone positioning, provided any meaningful contribution to predicting a successful next proning episode. CONCLUSIONS: In mechanically ventilated COVID-19 patients, predicting the success of prone positioning using clinically relevant and readily available parameters from electronic health records is currently not feasible. Given the current evidence base, a liberal approach to proning in all patients with severe COVID-19 ARDS is therefore justified and in particular regardless of previous results of proning.

Pulmonary Procoagulant and Innate Immune Responses in Critically Ill COVID-19 Patients.

Rationale: Systemic activation of procoagulant and inflammatory mechanisms has been implicated in the pathogenesis of COVID-19. Knowledge of activation of these host response pathways in the lung compartment of COVID-19 patients is limited. Objectives: To evaluate local and systemic activation of coagulation and interconnected inflammatory responses in critically ill COVID-19 patients with persistent acute respiratory distress syndrome. Methods: Paired bronchoalveolar lavage fluid and plasma samples were obtained from 17 patients with COVID-19 related persistent acute respiratory distress syndrome (mechanical ventilation > 7 days) 1 and 2 weeks after start mechanical ventilation and compared with 8 healthy controls. Thirty-four host response biomarkers stratified into five functional domains (coagulation, complement system, cytokines, chemokines and growth factors) were measured. Measurements and Main Results: In all patients, all functional domains were activated, especially in the bronchoalveolar compartment, with significantly increased levels of D-dimers, thrombin-antithrombin complexes, soluble tissue factor, C1-inhibitor antigen and activity levels, tissue type plasminogen activator, plasminogen activator inhibitor type I, soluble CD40 ligand and soluble P-selectin (coagulation), next to activation of C3bc and C4bc (complement) and multiple interrelated cytokines, chemokines and growth factors. In 10 patients in whom follow-up samples were obtained between 3 and 4 weeks after start mechanical ventilation many bronchoalveolar and plasma host response biomarkers had declined. Conclusions: Critically ill, ventilated patients with COVID-19 show strong responses relating to coagulation, the complement system, cytokines, chemokines and growth factors in the bronchoalveolar compartment. These results suggest a local pulmonary rather than a systemic procoagulant and inflammatory "storm" in severe COVID-19.

Ventilating two patients with one ventilator: technical setup and laboratory testing.

With a modified circuit, it is feasible to ventilate two patients with one ventilator over a relevant range of compliances. Adding inspiratory resistance allows individual titration of tidal volume, and incorporating one-way valves prevents pendelluft. https://bit.ly/3ex8SYP.