Predicting mortality from ARDS and COVID-19 infection: ‘Validation of severity-adjusted Berlin criteria at the height of Bogotá

Acute Respiratory Distress Syndrome (ARDS) is defined according to the Berlin criteria which include levels of Arterial Oxygen Pressure over Inspired Fraction of Oxygen (PaO2/FiO2) to define the severity of the these criteria were validated at sites with altitudes less than 1000 metres above sea level. than 1000 metres above sea level, the authors propose a formula for the adjustment of the PaO2/FiO2 at higher altitude sites, but this adjustment has not been validated in clinical studies. clinical studies, its behaviour and relationship with mortality is unknown.
Objective:
To establish PaO2/FiO2 at altitude in Bogotá (2626 masnm) for expected mortality according to severity of ARDS in patients with COVID-19.
a. Mild ARDS: PaO2/FiO2 for an expected mortality of 27% to 34.9%.
b. Moderate ARDS: PaO2/FiO2 for an expected mortality of 32% to 40.3%.
c. Severe ARDS: PaO2/FiO2 for an expected mortality 45% to 46.1%.
Methodology:
Analytical ambispective cohort study that will determine mortality for each level of severity using the formula recommended by the Berlin definition and subsequent sensitivity analysis of different cut-off points of the sensitivity analysis of different PaO2/FiO2 cutoff points to establish which oxygenation indices are associated with which oxygenation indices are associated with an expected mortality for each severity level. A mortality prediction model a mortality prediction model using the Operator Selection and Absolute Minimum Shrinkage method. Absolute Least Squares Shrinkage Operator (LASSO) to select the most relevant variables from the set of candidate predictors using the
candidate predictors using the logistic regression method and a classification and regression tree (CART) will also be used. (CART) will be used, using all collected variables in order to uncover effect modifiers and complex interactions. effect modifiers and complex interactions between the variables selected for the model. model.
Expected results:
To define the PaO2/FiO2 levels, which are associated with expected mortality according to the Berlin consensus for each severity level at the Bogotá altitude. Verify whether the recommended adjustment for PaO2/FiO2 for altitude is valid in clinical practice.

Respiratory distress, altitude, mortality