Mechanics of Breathing and Gas Exchange in Mechanically Ventilated Patients with COVID-19-associated Respiratory Failure.

To the Editor:

The acute lung insult resulting from severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has multifarious clinical presentations ranging from limited mild respiratory symptoms to a potentially fatal multifocal pneumonia/acute respiratory distress syndrome (ARDS) requiring weeks of mechanical ventilation. Whether these clinical presentations represent different levels of severity of the same “disease” or result from profoundly different pathophysiological mechanisms (virus invasion vs. inflammatory response of the host) remains an unanswered question. Three case series very recently published in the Journal (1–3) have reported conflicting data on the mechanical properties of the respiratory system and the gas-exchange profile observed in intubated patients presenting with SARS-CoV-2–induced respiratory failure. We have reanalyzed the data presented in these cases series (1–3) in an attempt to reconcile these discrepant observations and revisit some of the conclusions and clinical implications of these studies.

Do mechanically ventilated patients with COVID-19 pneumonia have well-preserved or deteriorated lung mechanics?

Gattinoni and colleagues (1) have reported in a cohort of 16 patients with a shunt fraction of ∼0.5, values of compliance of the respiratory system (Crs) averaging 50.2 ± 14.3 ml/cm H2O (1), that is, ∼60% from normal. Based on these observations, the authors concluded that a relatively preserved compliance in patients with coronavirus disease (COVID-19) pneumonia would make “high” positive end-expiratory pressure (PEEP) ineffective, and thus unnecessarily dangerous, and make prone position worthless because of a low benefit/resource ratio. However, Crs values in this study were exceptionally variable, ranging from 20 to 90 ml/cm H2O. In other words, a significant reduction in Crs is present in intubated patients with COVID-19, at least at some point during the evolution of the disease. Second, low Crs values averaging 35.7 ± 5.8 ml/cm H2O (in eight consecutive patients with COVID-19 studied at Day 1 after intubation) and 19.58 ± 7.96 ml/cm H2O (worst respiratory mechanics in 12 patients with COVID-19) were reported by Liu and colleagues (2) and by Pan and colleagues (3), respectively. Despite the claim of preserved elastic properties in COVID-19 pneumonia, these values of Crs are not very different from those reported in patients with ARDS (4, 5), as illustrated in Figure 1. To try to understand the discrepancy in Crs values between these studies and their variability, we have recomputed the individual data reported by Pan and colleagues (3) and found a significant correlation between the level of PEEP used in their patients and Crs (Figure 1A); PEEP levels were determined as the difference between the plateau pressure and the driving pressure. This surprising relationship implies that the lowest PEEP levels were used in patients with the lowest Crs and vice versa. For instance, a PEEP of 4 cm H2O was used in a patient with a Crs of 12 ml/cm H2O, whereas another patient with a Crs of 30 ml/cm H2O was exposed to a PEEP of 15 cm H2O. In addition, because a significant increase in alveolar Pco2 (PaCO) was always present as low Vt was used (3), we recomputed alveolar Po2 (PaO) based on the data available (3). PaO was calculated according to the alveolar gas equation using PaCO and FiO provided (3), and the gradient PaO–PaO was determined. These gradients were greatly deteriorated (Figure 1), as previously reported (1); yet, patients with the lowest compliance were also those with the highest PaO–PaO gradient (Figure 1). This indicates that despite an unusual severity of hypoxemia in this population, a coupling between low compliance and high arterial–alveolar O2 gradient is present in COVID-19–associated respiratory failure. This implies that “sufficient” levels of PEEP should be used in patients with COVID-19–associated respiratory failure and low Crs, as suggested by Figure 1. The optimal level of PEEP should be determined in any given patient by measuring Crs while increasing the PEEP level. Being able to shift the volume–pressure curve of the respiratory system to the right by using the appropriate PEEP may prove to be crucial in these patients. In any case, the levels of optimal PEEP should be determined in every individual patient with COVID-19–associated respiratory failure by considering the minimal level of end-expiratory pressure needed to decrease the driving pressure/volume ratio as shown in Figure 1.

Figure 1.

(A) Values of compliance (Crs) collected in mechanically ventilated patients with coronavirus disease (COVID-19) compared with data reported in acute respiratory distress syndrome (the references of the selected studies are given in the figure). Although data were not obtained at the same time of the disease, alterations of the elastic properties of the respiratory system can be significant in all these patients and are not dramatically different between patients with COVID-19 and acute respiratory distress syndrome. (B) Relationship between positive end-expiratory pressure (PEEP) and Crs, showing that when low levels of PEEP were used, low Crs was always present (see text for comments and discussion). (C) Crs versus PaO–alveolar PO (PaO) gradient. Extreme deterioration of the PaO–PaO gradient was observed in many patients; however, although the patients with the lowest Crs have the greatest gradient, the correlation remains weak in this limited population. (D) Relationship between the Crs/PEEP ratio and the PaO–PaO gradient; the ratio was used as an indicator of the effects of PEEP applied at any given Crs. The patients with the lowest ratio had the highest gradient, with a significant correlation between the two variables. (E) Iso–dead space (iso-Vd) curves showing the relationship between Vt and Vd/Vt ratio. By minimally increasing Vt, the change in Vd/Vt ratio and thus in alveolar gas composition improves out of proportion of the changes in serial Vd (see text for further comments). ARDS = acute respiratory distress syndrome.

Does minimally increasing Vt improve pulmonary gas exchange, or are the COVID-19 lungs nonrecruitable?

Lui and colleagues have shown that increasing Vt from 7 to 7.5 ml/kg produced a significant decrease in PaCO (2). We have reevaluated this question by determining the averaged dead space ventilation (V˙d) in patients receiving a Vt of 7 ml/kg (2). To do so, average alveolar ventilation (V˙a) was calculated from PaCO (V˙a = k × V˙co2/PaCO), and then V˙d was determined as V˙e (given in the text) minus V˙a. Based on the average body weight, Vt was computed and then f was determined from the e values, given in the text. The corresponding dead space (Vd) was computed as V˙d/f. The same computation was performed for a Vt of 7.5 ml/kg. The expected changes in Vd/Vt ratio were then calculated as a function of Vt (Figure 1) at the given Vds, creating iso-Vd curves. As shown in Figure 1, when Vt was increased from 7 to 7.5 ml/kg, the decrease in the Vd/Vt ratio was much higher than expected from a monoalveolar model (same iso-Vd curve), reflecting the recruitment of lung regions with high V˙a:Q˙ ratio (lowering Vd). These data therefore suggest that at a low “cost” in terms of barotrauma, it is possible via a modest increase in Vt to reduce serial dead space ventilation (as expected) together with a decrease in parallel dead space ventilation.

The phenotype of patients in acute respiratory failure with “COVID lungs” is certainly quite heterogenous; the individual determination of Crs, Pa–PaO gradient, and PaCO as a function of the level of PEEP and Vt should be performed in every patient to tailor the optimal modality of ventilation at the different stages of the disease. The short- and long-term impacts of using “larger” Vt together with relatively high PEEP in patients with COVID-19–associated respiratory failure who display a low compliance at low PEEP is fundamental to evaluate. Only such an approach could allow operation with the highest possible compliance and lowest Pa–PaO gradient in these patients.