Optimising respiratory support for early COVID-19 pneumonia: a computational modelling study

Prognostic analysis of high-flow nasal cannula therapy and non-invasive ventilation in mild to moderate hypoxemia patients and construction of a machine learning model for 48-h intubation prediction-a retrospective analysis of the MIMIC database

Background

This study aims to investigate the clinical outcome between high-flow nasal cannula (HFNC) and non-invasive ventilation (NIV) therapy in mild to moderate hypoxemic patients on the first ICU day and to develop a predictive model of 48-h intubation.

Methods

The study included adult patients from the MIMIC III and IV databases who first initiated HFNC or NIV therapy due to mild to moderate hypoxemia (100 < PaO2/FiO2 ≤ 300). The 48-h and 30-day intubation rates were compared using cross-sectional and survival analysis. Nine machine learning and six ensemble algorithms were deployed to construct the 48-h intubation predictive models, of which the optimal model was determined by its prediction accuracy. The top 10 risk and protective factors were identified using the Shapley interpretation algorithm.

Result

A total of 123,042 patients were screened, of which, 673 were from the MIMIC IV database for ventilation therapy comparison (HFNC n = 363, NIV n = 310) and 48-h intubation predictive model construction (training dataset n = 471, internal validation set n = 202) and 408 were from the MIMIC III database for external validation. The NIV group had a lower intubation rate (23.1% vs. 16.1%, p = 0.001), ICU 28-day mortality (18.5% vs. 11.6%, p = 0.014), and in-hospital mortality (19.6% vs. 11.9%, p = 0.007) compared to the HFNC group. Survival analysis showed that the total and 48-h intubation rates were not significantly different. The ensemble AdaBoost decision tree model (internal and external validation set AUROC 0.878, 0.726) had the best predictive accuracy performance. The model Shapley algorithm showed Sequential Organ Failure Assessment (SOFA), acute physiology scores (APSIII), the minimum and maximum lactate value as risk factors for early failure and age, the maximum PaCO2 and PH value, Glasgow Coma Scale (GCS), the minimum PaO2/FiO2 ratio, and PaO2 value as protective factors.

Conclusion

NIV was associated with lower intubation rate and ICU 28-day and in-hospital mortality. Further survival analysis reinforced that the effect of NIV on the intubation rate might partly be attributed to the other impact factors. The ensemble AdaBoost decision tree model may assist clinicians in making clinical decisions, and early organ function support to improve patients' SOFA, APSIII, GCS, PaCO2, PaO2, PH, PaO2/FiO2 ratio, and lactate values can reduce the early failure rate and improve patient prognosis.

An in-silico porcine model of phosgene-induced lung injury predicts clinically relevant benefits from application of continuous positive airway pressure up to 8 h post exposure

We present the first computational model of the pathophysiological consequences of phosgene-induced lung injury in porcine subjects. Data from experiments previously performed in several cohorts of large healthy juvenile female pigs (111 data points from 37 subjects), including individual arterial blood gas readings, respiratory rate and heart rate, were used to develop the computational model. Close matches are observed between model outputs (PaO2 and PaCO2) and the experimental data, for both terminally anaesthetised and conscious subjects. The model was applied to investigate the effectiveness of continuous positive airway pressure (CPAP) as a pre-hospital treatment method when treatment is initiated at different time points post exposure. The model predicts that clinically relevant benefits are obtained when 10 cmH2O CPAP is initiated within approximately 8 h after exposure. Supplying low-flow oxygen (40%) rather than medical air produced larger clinical benefits than applying higher CPAP pressure levels. This new model can be used as a tool for conducting investigations into ventilation strategies and pharmaceutical treatments for chemical lung injury of diverse aetiology, and for helping to refine and reduce the use of animals in future experimental studies.

Effects of CPAP and FiO2 on respiratory effort and lung stress in early COVID-19 pneumonia: a randomized, crossover study

BACKGROUND

in COVID-19 acute respiratory failure, the effects of CPAP and FiO2 on respiratory effort and lung stress are unclear. We hypothesize that, in the compliant lungs of early Sars-CoV-2 pneumonia, the application of positive pressure through Helmet-CPAP may not decrease respiratory effort, and rather worsen lung stress and oxygenation when compared to higher FiO2 delivered via oxygen masks.

METHODS

In this single-center (S.Luigi Gonzaga University-Hospital, Turin, Italy), randomized, crossover study, we included patients receiving Helmet-CPAP for early (< 48 h) COVID-19 pneumonia without additional cardiac or respiratory disease. Healthy subjects were included as controls. Participants were equipped with an esophageal catheter, a non-invasive cardiac output monitor, and an arterial catheter. The protocol consisted of a random sequence of non-rebreather mask (NRB), Helmet-CPAP (with variable positive pressure and FiO2) and Venturi mask (FiO2 0.5), each delivered for 20 min. Study outcomes were changes in respiratory effort (esophageal swing), total lung stress (dynamic + static transpulmonary pressure), gas-exchange and hemodynamics.

RESULTS

We enrolled 28 COVID-19 patients and 7 healthy controls. In all patients, respiratory effort increased from NRB to Helmet-CPAP (5.0 ± 3.7 vs 8.3 ± 3.9 cmH2O, p < 0.01). However, Helmet's pressure decreased by a comparable amount during inspiration (- 3.1 ± 1.0 cmH2O, p = 0.16), therefore dynamic stress remained stable (p = 0.97). Changes in static and total lung stress from NRB to Helmet-CPAP were overall not significant (p = 0.07 and p = 0.09, respectively), but showed high interpatient variability, ranging from - 4.5 to + 6.1 cmH2O, and from - 5.8 to + 5.7 cmH2O, respectively. All findings were confirmed in healthy subjects, except for an increase in dynamic stress (p < 0.01). PaO2 decreased from NRB to Helmet-CPAP with FiO2 0.5 (107 ± 55 vs 86 ± 30 mmHg, p < 0.01), irrespective of positive pressure levels (p = 0.64). Conversely, with Helmet's FiO2 0.9, PaO2 increased (p < 0.01), but oxygen delivery remained stable (p = 0.48) as cardiac output decreased (p = 0.02). When PaO2 fell below 60 mmHg with VM, respiratory effort increased proportionally (p < 0.01, r = 0.81).

CONCLUSIONS

In early COVID-19 pneumonia, Helmet-CPAP increases respiratory effort without altering dynamic stress, while the effects upon static and total stress are variable, requiring individual assessment. Oxygen masks with higher FiO2 provide better oxygenation with lower respiratory effort. Trial registration Retrospectively registered (13-May-2021): clinicaltrials.gov (NCT04885517), https://clinicaltrials.gov/ct2/show/NCT04885517 .

Simulation to minimise patient self-inflicted lung injury: are we almost there?

Computational modelling has been used to enlighten pathophysiological issues in patients with acute respiratory distress syndrome (ARDS) using a sophisticated, integrated cardiopulmonary model. COVID-19 ARDS is a pathophysiologically distinct entity characterised by dissociation between impairment in gas exchange and respiratory system mechanics, especially in the early stages of ARDS. Weaver and colleagues used computational modelling to elucidate factors contributing to generation of patient self-inflicted lung injury, and evaluated the effects of various spontaneous respiratory efforts with different oxygenation and ventilatory support modes. Their findings indicate that mechanical forces generated in the lung parenchyma are only counterbalanced when the respiratory support mode reduces the intensity of respiratory efforts.