COVID-19 Associated ARDS: ‘CARDS’ & the L-Phenotype
Jon-Emile S. Kenny MD [@heart_lung]
“No facts are to me sacred; none are profane; I simply experiment, an endless seeker with no Past at my back.”
-Ralph Waldo Emerson
When Drs. Marini and Gattinoni author a letter describing a new phenotype of the acute respiratory distress syndrome [ARDS] encountered in patients infected with SARS-CoV-2, we should all take a moment to read their words. In brief, an atypical hypoxemia is being observed – one in which calculated respiratory system compliance is strangely preserved. Thus, two phenotypes are born, the “L” [as in low] and “H” [as in high] subtypes of COVID-19 associated ARDS [i.e. ‘CARDS’]. While certainly noteworthy from a purely descriptive perspective, what is most controversial is that they propose using slightly lower PEEP [e.g. < 10 cm H2O] and higher tidal volumes [e.g. 7-9 mL/kg] when managing L-type CARDS. What are the key differences between L and H-type CARDS? And on what physiological principles are these management recommendations based?
Early in CARDS, there is significant hypoxemia, but with relatively preserved respiratory system elastance [elastance is the inverse of compliance, so a low elastance is mathematically the same as a high compliance]. Importantly, in typical, severe ARDS, the specific stiffness, or specific elastance, of the lung does not change appreciably. In humans, the specific elastance of the lungs – even in the throes of horrific, classical ARDS – stays constant at roughly 12 cm H2O. How is this possible? The key is that the specific elastance relates lung strain to lung stress and that the lung strain is the change in lung volume relative to its resting volume. Thus, if the functional residual capacity (FRC) is 500 mL, and the tidal volume delivered is 500 mL, then the lung strain is 1.0 and its stress is 12 cm H2O. Conversely, if the FRC is 167 mL [i.e. a ‘baby lung’] and the tidal volume delivered is 500 mL, then the lung strain is 3.0 and its stress, therefore, is 36 cmH2O; the stress is the pressure across the lung. In both patients, the tidal volume is the same and the specific elastance is the same, but the pressure generated is 3 times larger in the second patient because the ‘baby lung’ is three times smaller. Clearly, the second patient would have a higher calculated respiratory system elastance than the first – suggesting that the lung is stiff, when in reality it is small.
In typical ARDS, there is a reasonably good relationship between the degree of hypoxemia and the shrinking baby lung such that as oxygenation worsens, strain, stress and calculated elastance are all also high [i.e. compliance falls]. With these abnormal mechanics, patients often breathe with a shallow pattern to minimize the elastic work, but rapidly to maintain carbon dioxide excretion. Unfortunately, a rapid-but-shallow breathing pattern itself can cause dead space and energetic failure of the respiratory pump.
But CARDS is not typical ARDS! There is a mismatch between the severity of hypoxemia and the calculated elastance. Given the above, we deduce that there is preserved elastance because the resting lung volume is close to normal and therefore lung strain and stress are near-normal. Thus, patients are less disposed to rapid-shallow-breathe and, consequently, may have less energetic demand placed upon the respiratory pump. CT evidence supports these findings; the early CARDS lung is seen to have low elastance, low lung weight, low response to positive end-expiratory pressure [PEEP] – hence the term ‘L’ type CARDS. The hypoxemia appears to be primarily a vascular event; the lungs have lost their protective ability to pinch off perfusion to areas of compromised gas exchange. It is not entirely known why this occurs, but imbalance between Angiotensin II and Angiotensin I-7 has been proposed as well as pulmonary vascular thromboses.
If COVID-associated ARDS progresses, the peripheral ground glass and interstitial edema which typify the L-type can morph into a typical ARDS pattern with dependent consolidations and ‘baby lung’ physiology. Then, the patient is observed to have high elastance [stiff lungs], high lung weight and high response to PEEP – coined ‘H’ type. In one series, the ‘H’ type CARDS included 20-30% of all patients.
Mechanical Ventilation for L-Type
Drs. Marini and Gattinoni reason that because L-type CARDS lacks lung to unfurl, the provision of PEEP cannot recruit and homogenize the lung. Thus, PEEP only begets harmful hemodynamics. For example, augmenting afterload in good tissue and, in doing so, feeding gasless lung deoxygenated blood. Further, as described previously, this maldistribution of blood flow can adversely affect the pulmonary vasculature [i.e. especially ‘corner vessels’], leading to stress fracture and worsening edema. So, when these patients are invasively ventilated, we are asked to use slightly lower PEEP with slightly higher tidal volumes; in other words, break the sacrosanct tidal volume recommendations for patients with classic ARDS or ‘H’ type CARDS.
The Lung is Also Viscous
I must admit, I was surprised by their invasive, mechanical ventilation strategies for the ‘L’ phenotype of CARDS. I was surprised because the linear relation between strain and stress, described above, applies to elastic tissue; but the lung is also viscous! The stress felt across viscous tissue is not proportional to strain, but rather strain velocity. This biophysical fact explains previous animal data [from Dr. Gattinoni’s group] showing that when total strain is held constant, but composed of different PEEP fractions [e.g. 100% tidal strain versus 75% tidal strain and 25% PEEP strain], there is more VILI in normal lungs when the PEEP fraction is less. Why? Because greater tidal strain means a larger volume change per unit time [i.e. heightened strain velocity] and amplified viscous stress across the lung.
And why wouldn’t the ground glass opacities found early in L-type CARDS act as stress raisers [lecture 5]? Gattinoni’s research has shown that the VILI vortex begins around elastic inhomogeneities in the normal porcine lung [e.g. at the interface of the lung parenchyma and an airway, pleural lining or blood vessel]. Could not border regions between ground glass opacities and normal lung intensify the viscous stress of higher tidal volume, lower PEEP ventilation? Could this not also be felt regionally by corner vessels and exacerbate vascular lung injury? The Meade model predicts that between normal lung and a stress raiser [like ground glass] viscoelastic stress is amplified by a factor of 4.5. Clinically, this has found this to be closer to a factor of 2.0.
And Driving Power?
And minimizing tidal strain rate about viscoelastic stress raisers is corroborated in other avenues of ARDS research such as ergotrauma and driving power. Previously, Gattinoni’s group found that both tidal volume and flow rate augment total mechanical power exponentially by a factor of 2.0, while respiratory rate does so by a factor of 1.4. Would not increasing tidal volume to decrease respiratory rate mathematically favour ergotrauma?
More Questions Than Answers
That COVID-19 presents an atypical lung injury typified by disproportionately low oxygenation relative to functional lung size seems incontrovertible. Without a ‘baby lung’, elastic breathing work is smaller and patients seem less inclined to rapid-shallow-breathe. Nonetheless, the L-phenotype comprises stress raisers able to amplify the viscous stress of high tidal strain rates! Can we ignore this and adopt a lower PEEP, higher tidal volume strategy? Or should we continue to minimize tidal strain velocity with traditional strategies? The only clarity for me is that no current arguments are sacred, none profane. And that the answer demands endless seekers and experimenters with respiratory physiology at their backs.
Dr. Kenny is the cofounder and Chief Medical Officer of Flosonics Medical; he is also the creator and author of a free hemodynamic curriculum at heart-lung.org