Effect of End-inspiratory Airway Pressure Measurements on the Risk of VILI in Ventilated Patients

Overview

Mechanical ventilation may be associated with ventilator-induced lung injury (VILI). Several respiratory variables have been employed to estimate the risk of VILI, such as tidal volumes, plateau pressure, driving pressure, and mechanical power. This dissipation of energy during ventilation can contribute to VILI through two mechanisms, stress relaxation and pendelluft, which can be estimated at the bedside by applying an end-inspiratory pause and evaluating the slow decrease in airway pressure going from the pressure corresponding to zero flow (called pressure P1) and the final pressure at the end of the pause (called plateau pressure P2). The choice of measuring the end-inspiratory airway pressure (PawEND-INSP) at a fixed, although relatively early, timepoint, i.e., after 0.5 second from the beginning of the pause, as prescribed by the indications of the Acute Respiratory Distress Syndrome (ARDS) Network, while assessing the risk of VILI associated with the elastic pressure of the respiratory system, may not reflect the harmful potential associated with the viscoelastic properties of the respiratory system. It is still unclear whether an PawEND-INSP measured at the exact moment of zero flow (P1) is more reliable in the calculation of those variables, such as ΔP and MP, associated with the outcomes of patients with and without ARDS, as compared to the pressure measured at the end of the end-inspiratory pause (plateau pressure P2). This multicenter prospective observational study aims to evaluate whether the use of P1, as compared to P2, affects the calculation of ΔP and MP. The secondary objectives are: 1) verify whether in patients with a lung parenchyma characterized by greater parenchymal heterogeneity, as assessed by EIT, P1-P2 decay is greater than in patients with greater parenchymal homogeneity; 2) evaluate whether patients with both ΔP values calculated using P1 and P2 \<15 cmH2O (or both MP values calculated using P1 and P2 \<17 J/min) develop shorter duration of invasive mechanical ventilation, shorter ICU and hospital length of stay and lower ICU and hospital mortality, as compared to patients with only ΔP calculated with P1 ≥ 15 cmH2O (or only MP calculated with P1 ≥ 17 J/min) and patients with both ΔP values calculated using P1 and P2 ≥ 15 cmH2O (or both MP values calculated using P1 and P2 ≥ 17 J/min).

Introduction Mechanical ventilation is necessary to ensure survival in critically ill patients but may be associated with ventilator-induced lung injury (VILI). Several respiratory variables have been employed to estimate the risk of VILI: * In patients with the acute respiratory distress syndrome (ARDS), the prevention of VILI by limiting ventilatory pressures and volumes (i.e., tidal volumes of 4-8 mL per kilogram of ideal body weight and plateau pressure \[Pplat\], i.e., the pressure measured in the respiratory system during an end-inspiration pause of 0.5 second in the absence of flow and correlated to the end-inspiratory alveolar damage, less than 30 cmH2O) has been shown to improve patient survival. These ventilation settings are surrogates, yet not ideal predictors, of the risk of volutrauma and barotrauma, respectively. * More recently, an important role in the pathogenesis of VILI has been attributed to the difference between Pplat and total positive end-expiratory pressure (PEEP), the latter measured after an end-expiratory pause. This difference, known as driving pressure (ΔP), represents the change in pressure above PEEP required to obtain the tidal volume in a patient without spontaneous inspiratory efforts and aggregate information about volutrauma, barotrauma, and atelectrauma in a single variable that is easily measurable at the bedside. The increase in ΔP has been associated with worse outcomes in patients with and without ARDS. Some observational studies have identified the ΔP value of 15 cmH2O as the threshold beyond which the risk of mortality in patients with ARDS increases significantly, but others have shown that a safety threshold value for ΔP cannot be identified, thus suggesting that the lower the ΔP, the lower the risk of mortality. * Some studies suggest that inspiratory flow and strain rate may contribute to VILI in experimental animals and patients with moderate-to-severe ARDS. Therefore, taking into account inspiratory flow may give more insight into the risk of VILI in mechanically ventilated patients. Mechanical power (MP) is the total energy transferred from the ventilator to the lungs during inspiration and includes variables such as inspiratory flow and respiratory rate. MP has been shown to predict mortality in patients with and without ARDS. * During ventilation, the lung has a viscoelastic mechanical behavior, dissipating energy both during inspiration and during expiration. This dissipation of energy can contribute to VILI through two mechanisms: stress relaxation, i.e., the release of the parenchymal tension accumulated during inspiration, and pendelluft, i.e., the redistribution of the volume of ventilation to the alveoli with a longer time constant. In patients with ARDS, the lung parenchyma is heterogeneous because of the coexistence of aerated alveolar units and other regions that are filled with edema or collapsed due to the superimposed pressure. Therefore, both the alveolar units with shorter time constants and those with longer time constants are at risk of VILI: the former are affected by higher transpulmonary pressures, while the latter are exposed to pendelluft. Even lungs of patients without ARDS may be heterogeneous and thus predisposed to VILI through these mechanisms, e.g., because of alveolar atelectasis or consolidation and bronchiolar obstruction. Some imaging techniques available for clinical use may help assessing the degree of lung parenchymal heterogeneity. Electrical impedance tomography (EIT) is a non-invasive bedside technique that allows assessing the distribution of lung ventilation and perfusion by recording the impedance variation to small electrical currents delivered by an electrode belt wrapped around the patient's chest. This method has been shown to visualize and measure pendelluft during controlled mechanical ventilation. The EIT variables used to assess lung heterogeneity include the center of ventilation, i.e., the variation in the distribution of ventilation according to a ventro-dorsal gradient, the global inhomogeneity index, an index that estimates the heterogeneity of ventilation, and the regional ventilation delay, indicating the delay in ventilation compared to global ventilation due to atelectrauma or differences in the time constant of different lung areas. Stress relaxation and pendelluft can be estimated at the bedside by applying an end-inspiratory pause. During this maneuver, the airway pressure curve exhibits two subsequent phases of decrease. First, a rapid pressure drop occurs, ranging from peak airway pressure to the pressure corresponding to zero flow (called pressure P1), which reflects the dissipation of pressure in the conduction airways. Then, a slow decrease in airway pressure follows, which goes from P1 to the final pressure at the end of the pause (called plateau pressure P2). The difference between P1 and P2 (P1-P2 decay) depends on stress relaxation and pendelluft and can be used as index of the time constant inequalities and viscoelastic tissue properties of the respiratory system. The choice of measuring the end-inspiratory airway pressure (PawEND-INSP) at a fixed, although relatively early, timepoint, i.e., after 0.5 second from the beginning of the pause, as prescribed by the indications of the ARDS Network, while assessing the risk of VILI associated with the elastic pressure of the respiratory system, may not reflect the harmful potential associated with the viscoelastic properties of the respiratory system. Previous pilot studies discouraged the use of inspiratory pauses of less than 3 seconds not to underestimate respiratory system compliance and resistance. However, a delayed measurement of the PawEND-INSP could lead to neglect the contribution of pulmonary viscoelastic properties and pendelluft to VILI. It is still unclear whether an PawEND-INSP measured at the exact moment of zero flow (P1) is more reliable in the calculation of those variables, such as ΔP and MP, associated with the outcomes of patients with and without ARDS, as compared to the pressure measured at the end of the end-inspiratory pause (plateau pressure P2). Rationale of the study * The effect of P1-P2 decay on the calculation of ΔP and MP are still unclear; * The use of P1 as PawEND-INSP can allow taking into account the risk of VILI associated with stress relaxation and pendelluft, potentially neglected by P2; * The identification of a different effect of the use of P1 or P2 on the calculation of ΔP and MP can help identifying lung protective ventilation settings in mechanically ventilated patients. Objectives of the study This multicenter prospective observational study aims to evaluate whether the use of P1, as compared to P2, affects the calculation of ΔP and MP. The secondary objectives are: 1) verify whether in patients with a lung parenchyma characterized by greater parenchymal heterogeneity, as assessed by EIT, P1-P2 decay is greater than in patients with greater parenchymal homogeneity; 2) evaluate whether patients with both ΔP values calculated using P1 and P2 \<1 5 cmH2O (or both MP values calculated using P1 and P2 \< 17 cmH2O) develop shorter duration of invasive mechanical ventilation, shorter ICU and hospital length of stay and lower ICU and hospital mortality, as compared to patients with only ΔP calculated with P1 ≥ 15 cmH2O (or only MP calculated with P1 ≥ 17 cmH2O) and patients with both ΔP values calculated using P1 and P2 ≥ 15 cmH2O (or both MP values calculated using P1 and P2 ≥ 17 cmH2O). Data collection Data will be collected for 1 year. Demographic, anthropometric and anamnestic variables, ventilation settings, respiratory mechanics, and EIT variables normally collected during daily clinical practice at the involved centers will be recorded within 48 hours since ICU admission. Data collection for each patient will end upon discharge from the hospital. The methodology of the study involves the following phases: 1. Measurement of the endotracheal cuff pressure with a manometer and adjustment to normal values 2. Suggested mechanical ventilation settings: switching to volume-controlled mode with 10% automatic inspiratory pause, 0% (or 0 seconds) inspiratory rise time, 1:2 inspiratory/expiratory ratio (or ratio of inspiratory time to total respiratory cycle time of 33%), tidal volume of 6 mL/kg of ideal body weight and patient with no spontaneous respiratory effors. Respiratory rate, PEEP and inspired oxygen fraction are set according to clinical indications. 3. First clip recording 4. 5-second end-inspiratory pause after one complete respiratory cycle has occurred (to include one complete respiratory cycle with no pause in the clip) 5. In the recorded clip, detection with a cursor of the following pressure values: 5.1. Peak pressure 5.2. Ppause: Pplat automatically recorded by the ventilator with the 10% automatic inspiratory pause (this can be also recorded during tidal breathing) 5.3. P1: after the initial brief fluctuation of the flow, P1 is read at the point of zero flow at the bottom of any pressure fluctuation caused by cardiac contraction 5.4. P2: P2 is recorded after 5 seconds of end-inspiratory pause at the bottom of any pressure fluctuation caused by cardiac contraction 5.5. P0.5s: PawEND-INSP recorded after 0.5 seconds during the pause at the bottom of any pressure fluctuation caused by cardiac contraction 5.6. P2s: PawEND-INSP recorded after 2 seconds during the pause at the bottom of any pressure fluctuation caused by cardiac contraction 5.7. P3s: PawEND-INSP recorded after 3 seconds during the pause at the bottom of any pressure fluctuation caused by cardiac contraction 5.8. Tidal volume at the end of the inspiratory hold to check for the presence of air leaks If the flow does not reach zero and/or the pressure waveform shows oscillations not attributable to cardiac contractions, but to the patient's respiratory efforts, the measurement is not reliable and cannot be recorded. The patient should be evaluated later on or excluded from the study. 6. Wait for 10 respiratory cycles to be completed 7. Second clip recording 8. 5-second end-expiratory pause after one complete respiratory cycle has occurred (to include one complete respiratory cycle with no pause in the clip) 9. In the second recorded clip, detection with the cursor of the following pressure values: 9.1. Increase in airway pressure from the value immediately preceding the occlusion and the plateau after 5 seconds (static intrinsic PEEP) 9.2. Increase in airway pressure from the end-expiratory value to the value at zero flow in the respiratory cycle without expiratory pause (dynamic intrinsic PEEP) If the flow does not reach zero and/or the pressure waveform shows oscillations not attributable to cardiac contractions, but to the patient's respiratory efforts, the measurement is not reliable and cannot be recorded. The patient should be evaluated later on or excluded from the study. 10. Check for the presence of airway closure (in patients without respiratory efforts): reduce respiratory rate to 6 breaths per minute, set inspiratory flow to 5 L/min, record the first breath, and check for the presence of an inflection point in the initial part of the pressure-time waveform not attributable to intrinsic PEEP. In case of airway closure, record the airway opening pressure value and ΔP will be calculated as the difference between PawEND-INSP and airway opening pressure (AOP). Data will be generated in the participating centers and recorded via web application on the servers of the University of Padua using the Research Electronic Data Capture (REDCap) management software developed by the Unit of Biostatistics, Epidemiology and Public Health of the University of Padua, which will be disseminated at a multicenter level.