Bedside calculation of mechanical power during volume- and pressure-controlled mechanical ventilation

Patients admitted to the intensive care unit from January to June 2018 after elective surgery (excluding thoracic and cranial surgery) or with medical disease were considered eligible. The study was approved by the institutional review board of our hospital (Comitato Etico Interaziendale Milano Area A 1, n. 628), and informed consent was obtained according to Italian regulations. Exclusion criteria were hemodynamic instability (systolic arterial pressure < 100 mmHg) and COPD (Gold stages 3 and 4).

Critically ill patients with acute respiratory failure, kept in supine position and sedated and paralyzed for clinical reasons, were ventilated in volume-controlled mode with a tidal volume of 6–8 mL/kg of predicted body weight with two inspiratory flows (high, 60 L/min, and low, 30 L/min), with a square waveform. Subsequently, pressure-controlled ventilation was applied to reach the same tidal volume as volume-controlled ventilation. PEEP and respiratory rates were not changed throughout the study.

The inspiratory flow rate was measured with a heated pneumotachograph (Fleisch no. 2, Fleisch, Lausanne, Switzerland). Airway pressure was measured proximally to the endotracheal tube with a dedicated pressure transducer (MPX 2010 DP, Motorola, Solna, Sweden). Flow and airway pressure were collected at a sampling rate of 100 Hz and stored for analysis (Colligo, Elekton, Milan, Italy).

The following variables were measured in each condition: tidal volume, inspiratory time, peak airway pressure, plateau airway pressure, and PEEP during an end-inspiratory and expiratory pause. Means were computed over five consecutive breaths.

The MP was calculated with a comprehensive and a surrogate formula.

where ΔP_insp is the pressure (cmH₂O) above PEEP during pressure-controlled ventilation, T_insp is the inspiratory time (s), and C and R are the respiratory system compliance (ml/cmH₂O) and resistances (cmH₂O s l⁻¹), respectively. Compliance and respiratory system resistance (airways and tissue resistances) were calculated during end-expiratory and end-inspiratory pauses from the data when the patient was ventilated in volume-controlled mode with an inspiratory flow of 30 L/min. The main assumption in this equation is that airway resistances are considered constant throughout the breath, which is not necessarily true, as resistances are also a function of flow, which varies during pressure-controlled ventilation.

The regressions between the MP assessed by the geometric method and the comprehensive algebraic formula for 30 and 60 L/min of inspiratory flow are shown in Figure S3. The two methods were closely correlated (r² = 0.96, p < 0.001, and r² = 0.97, p < 0.001). The Bland-Altman analyses are reported in Fig. 1. For an inspiratory flow of 30 L/min, the mean difference was − 0.053 J/min with upper and lower limits of agreement (± 1.96 SD) of 0.77 and − 0.81 J/min; for an inspiratory flow of 60 L/min, the mean difference was − 0.4 J/min with upper and lower limits of agreement of 0.7 and − 1.5 J/min. The MP calculated with the surrogate equation was also closely correlated (r² = 0.95 and 0.94) with bias of − 0.0074 (0.91, − 0.93) and − 1.0 (0.45, − 2.52) J/min at low and high inspiratory flows (Fig. 2 and S4).

The regression between the MP computed by the geometric method and comprehensive algebraic formula is shown in Figure S5, left panel. The two methods were closely correlated (r² = 0.81, p < 0.001). The Bland-Altman analyses are reported in Fig. 3a. The mean difference was − 0.001 J/min, with upper and lower limits of agreement 2.05 and − 2.05 J/min.

The MP with the surrogate equation were closely correlated (r² = 0.94, p < 0.001) with a bias of − 0.81 J/min (2.11, − 0.48 upper and lower limits of agreement) (Fig. 3b and S5, Panel B).

The mean values in each experimental set-up are summarized in Table S1.

For volume-controlled ventilation, Gattinoni et al. originally proposed an equation that was complicated to compute at the bedside because it included all the respiratory physiological variables [10]. The same authors proposed also a simpler equation that, starting from the measured values of peak, plateau and end-expiratory pressure was mathematically identical to the original one, allowing to calculate MP with the data displayed by the ventilator, provided that an end-inspiratory hold was performed. We called this “comprehensive formula.” Recently, Giosa and colleagues [18] proposed a surrogate formula that allowed to calculate MP without the inspiratory hold. In their original paper, however, the equation was not tested against the reference geometrical method. The present study indicates that for volume-controlled ventilation, the MP measured by the comprehensive and the surrogate equations were highly correlated with the reference-standard geometric method, with limits of agreement mostly within 2 J/min. This advocates that the simpler surrogate formula is of high clinical utility, allowing to quickly calculate MP even in settings where performing an inspiratory hold is either not possible (operating theater, patient transport) or not feasible (elevated number of patients and/or staff shortage). Of note, the fixed value for resistances assumed by the surrogate equation quite likely underestimates the “real” MP at high inspiratory flow rates, as the calculated resistances tend to increase because of the turbulent motion component. However, Giosa et al. found that even doubling the resistance up to 20 cmH₂O s/L, the underestimation bias remained quite low—in the order of 1.3 J/min. In the current study, to assess this surrogate method at two different resistances (high and low) directly against the geometric method, we applied high and low inspiratory flows (30 and 60 L/min). At both levels, the error between the geometric and the surrogate methods remained clinically negligible.

As shown in Table S1, both the measured (geometrical method) and the calculated (comprehensive and surrogate formulas) MP at 60 L/min was higher than at 30 L/min. This is not unexpected: as already stated by Gattinoni et al. in the original paper [10], MP increases with increasing airway resistances and flow. Indeed, at higher flows, not only the airway resistances, but also the tissue resistances increase, as the higher rate of tidal strain requires more energy. While it can be reasonably assumed that the energy dissipated in the airways and the endotracheal tube does not cause VILI, the one dissipated to perform the tissue strain has been already shown to cause damage and should not be ignored [23].

Until Becher et al. [19] and Van der Meijden et al. [20] reported two equations to measure the MP in pressure-controlled ventilation, it had been erroneously assessed with equations developed and validated only for volume-controlled ventilation. For example, Serpa Neto et al. [15] computed the MP in a big population of mechanically ventilated patients using a single formula independently of the actual type of ventilation. Van der Meijden et al. finally suggested a comprehensive formula that considers all the variables. The main drawback of this approach is the flow-dependency of the resistances: as flow is not constant in pressure-controlled ventilation, whatever value used is intrinsically approximated. Nevertheless, the need of an inspiratory hold, calculation of the resistances, and the presence of an exponential make this equation usable only if directly implemented in the ventilator. Becher et al. besides their original equation proposed an extremely easy to use surrogate formula for the same purpose. Interestingly, in our population, the surrogate performed equally if not better than the comprehensive formula from Van der Meijden, having a higher R² and, despite a trivially larger bias, a smaller confidence interval. This discrepancy between the comprehensive and the surrogate formulas is probably due to the assumptions that need to be made to calculate the resistances during an inflation with decreasing flow. The clinical consequence is that the surrogate formula, despite its simplicity, has a high degree of accuracy and can be safely used at the bedside without the need of complex calculations.

Nowadays, plateau pressure and driving pressure are the de facto clinically used variables to monitor whether a certain mechanical ventilation setup is protective or not [6]. The main benefit of driving pressure is the extreme ease of the calculation that, together with its association with mortality, allowed a rapid adoption around the world. Nevertheless, we believe that mechanical power, also including the effects of respiratory rate, flow, and PEEP, gives a much more comprehensive view on the ventilator-related causes of VILI [24] but its widespread application has been hindered by the complexity of its calculation. We believe that our data do support the use of the surrogate formulas at the bedside to add a tile in the puzzle of lung protection. Clinicians are used to use low tidal volumes (6 mL/kg, driving pressure < 14 cmH₂O), and to maintain a viable CO₂ clearance, they ramp up the respiratory frequency. It is therefore of paramount importance to sensitize the intensivists to consider that high respiratory rate does injure the lung as it is part of the whole energy package delivered by the ventilator [11, 25].

This study has several limitations. First, the comprehensive and surrogate equations can be currently applied only in sedated and paralyzed patients with no active breathing. Second, it is reasonable to think that the biological effects of MP will be clearer once: (1) MP is partitioned to the lung only by using the transpulmonary pressure [26] instead of the total airway pressure and (2) it is normalized to the lung volume or to the amount of well-aerated lung tissue, as MP is an extensive property and might have different effects depending on the amount of lung mass bearing a given energy load [14].