Respiratory therapy in COVID-19: Which model?
Short title: COVID-19: Which model?
Author Information:
Amit Jain M.D., Anesthesiology Institute, Cleveland Clinic Abu Dhabi, United Arab Emirates
D. John Doyle, M.D., Ph.D., Anesthesiology Institute, Cleveland Clinic Abu Dhabi, United Arab Emirates
Corresponding Author:
Amit Jain
Anesthesiology Institute, Cleveland Clinic Abu Dhabi, Al Maryah Island, Abu Dhabi
United Arab Emirates
M: +971544103790
Email:amitvasujain@gmail.com
Word and Element Counts:
Word Count: 1232
Number of references: 10
Number of figures: 1
Funding Statement: Support was provided solely from institutional and/or departmental sources
Conflicts of Interest: The authors declare no competing interests
Dear Editor,
As we deal with the COVID-19 pandemic we face a controversy concerning its pathophysiology and how to integrate available knowledge into practice while awaiting study outcomes. COVID-19’s pathophysiology remains elusive, as reflected in putative mechanisms that remain unsupported by robust evidence. Some models draw on clinical observations without reference to supporting data from genomic, proteomic, molecular, physiological, and other data pertaining to human coronaviruses. Consequently, some proposed models for COVID-19 pathophysiology and their corresponding treatment options remain highly divergent.
A key example is the debate over whether COVID-19 pneumonia presents with typical ARDS features. Since COVID-19 pneumonia patients frequently have hypoxemia that fulfils the Berlin definition for ARDS early in the disease, these patients are often ventilated using high PEEP-low tidal volume strategies, with early intubation being recommended to reduce aerosol-based viral spread. However, of necessity, both these recommendations were based purely on clinical judgment without reference to the underlying pathophysiology (which was lacking at the time of these recommendations) or the outcomes of clinical trials (which were just being launched). Consequently, such recommendations sometimes caused more harm than benefit.
As discussed below, various theories have been proposed but none explain the disease pathophysiology in a satisfactory manner.
1. The Gattinoni Model
Gattinoni et al. suggested two phenotypes for COVID-19 pneumonia: “Type-L” and “Type-H”[1], where patients present initially with the “L-phenotype” characterized by severe hypoxia despite normal lung compliance and low recruitability. Patient self-inflicted lung injury (P-SILI) and high PEEP injury is said to cause progression to the “H-phenotype”, similar in pathology and treatment to ARDS. They suggested that high PEEP / low tidal volume ventilation protocols are ineffective in Type-L patients and could even cause harm. After noting that the ratio of shunt fraction to fraction of gasless tissue on CT scan averaged 3.0 ± 2, suggesting hyperperfusion of gasless tissue[2], the authors additionally hypothesized that loss of lung perfusion regulation and vasoplegia caused the hypoxia in Type-L patients.[1]
Gattinoni et al. later suggested endothelial dysfunction as the reason for pulmonary vasoplegia in Type-L pneumonia [3], although no specific mechanism was proposed. Still, the theory received wide recognition and become a basis for ventilatory management in many ICUs worldwide.
2. High altitude pulmonary edema (HAPE)
Some physicians compare COVID-19 symptoms to those found in HAPE, offering pulmonary vasoconstriction in similarity to HAPE as the underlying mechanism for hypoxia in Type-L patients.[4] Even though this hypothesis is unsupported by evidence, the authors recommended acetazolamide, nifedipine and phosphodiesterase inhibitors as possible treatment options. Understandably, clinicians with experience with both HAPE and COVID-19 have pushed back on this observation and have argued that the comparison is risky. [5]
Overall, the early clinical presentation of COVID-19 disease is very confusing (especially silent hypoxia) despite the fact that the SARS-CoV-2 has 80% genomic similarity with SARS CoV-1. However, most of the late manifestation of the COVID-19 disease including, ARDS as well as the clinico-pathological findings are very similar to SARS and MERS. Based on these observations, we speculate that the main difference lies in the reduced virulence and immunogenicity of SARS-CoV-2 in comparison to SARS-CoV-1; thus, preventing direct pulmonary epithelial cell injury and thereby, delaying the progression to ARDS (median 8 days after the onset of symptoms) than the more usual manner as seen in other pulmonary infections. This hypothesis is supported by the preliminary cell culture study from Wuhan by Zhu et al. [6]. On human airway epithelial cells, only mild cytopathic effects in the form of loss of cilium beating and no cell lysis were observed 96 hours after viral inoculation. No specific cytopathic effects were observed in the Vero E6 and Huh-7 cell lines until 6 days after inoculation. Recently, Chu H et al. compared the replication kinetics and cell damage profiling of SARS-CoV-2 and SARS-CoV in non-human primate cells. [7] SARS-CoV2 consistently induced significantly delayed and milder levels of cell damage than did SARS-CoV, but with high replicability, supporting high transmissibility. [7] SARSCoV- 2 also exhibited more efficient replication but, induced significantly less host interferon and proinflammatory response than SARS-CoV ex vivo human lung tissue. [8]
To provide a pathophysiological model that better describes such a clinical course as well as different phenotypic presentations of the COVID-19 disease in concordance with the existing research on the renin-angiotensin system, previously described pathophysiological processes for other human coronavirus infections and the genomic similarities between the SARS and SARS-CoV-2 viruses, and weak cytolytic properties of SARS-CoV-2, we developed a conceptual model that we believe can help to explain the pathogenesis of COVID-19.
‘Epithelial-endothelial crosstalk’ hypothesis
In our view the L-phenotype is actually a Stage 2b/Stage 3 manifestation of COVID-19 pneumonia [9] and we offer an ‘epithelial-endothelial crosstalk’ hypothesis involving alterations in local RAS system as a mechanistic explanation:
1. SARS-CoV-2 enters type-2 pneumocytes following binding to ACE2 receptors, causing downregulation of ACE2 on the alveolar-capillary membrane (ACM). This reduces the level of protective, vasodilatory, anti-proliferation and anti-inflammatory ACE2-Ang 1-7-masR activity and increases the level of vasoconstriction, proliferation, inflammation and pro-fibrotic ACE 1-Ang II-AT1-R activity.
2. Angiotensin II mediated activation of alveolar endothelium rich in AT1 receptors increases the release of endothelin-1 (ET1) and reactive oxygen species (ROS). Meanwhile, ACE2-Ang 1-7-masR mediated constitutive endothelial nitric oxide synthase (eNOS) activation and NO release is relatively inhibited (Figure 1). Of note, SARS-CoV2 seems to produce little cytopathic changes in pulmonary epithelium.[7]
3. Intense ROS-mediated and ET1-mediated pulmonary vasoconstriction is hypothesized to be the principal mechanism for the initial hypoxia. This vasoconstriction is severe but, uneven, and as a result, capillary beds with relatively less vasoconstriction are disproportionately exposed to elevated microvascular pressures, resulting in recruitment and regional over-perfusion. This results in increased shunt fraction and hypoxia. As the disruption of the ACM progresses, proteins, fibrin, cells, and fluid leak into the alveolar space, resulting in bilateral patchy ground glass opacities on CT scan. Additionally, the finding of elevated pulmonary artery pressure, an enlarged right-atrium, and diminished right-ventricle (RV) function in patients with P/F ratios ≥ 150 mmHg [10] suggest pulmonary vasoconstriction as a mechanism for RV longitudinal strain (RVLS). Notably, RVLS has been linked with increased risk for progression to ARDS and mortality in COVID-19.
4. Following ACM disruption, SARS CoV-2 enters the pulmonary capillaries and infects the pulmonary endothelial cells via ACE-2 protein on the luminal surfaces. As a result, endothelial cells assume a ‘proinflammatory’/ ‘procoagulant’ phenotype, causing thrombotic occlusions of heterogeneous regions of pulmonary vasculature.
5. Activated endothelial cells accelerate apoptosis of alveolar epithelial and endothelial cells and produce a cytokine storm syndrome (CSS).
6. Hematogenous spread resulting in direct vial-induced injury and CSS can explain extrapulmonary disease manifestations.
In sum, our hypothesis supports the observations of Solaimanzadeh and Gattinoni et al. but suggest an alternate pathophysiology pathway common to SARS and MERS. In our opinion, the stratification of COVID-19 patients using the Berlin criteria is problematic. Finally, our hypothesis has treatment implications warranting investigation. Ventilation strategies as proposed by Marini et al.[3] may be effective but whether early intubation can prevent a transition to ARDS is difficult to explain based on our epithelial-endothelial cross-talk hypothesis. An RV protective approach including early prone positioning and inhaled NO, as well as steroids and anticoagulation may prove useful. Human recombinant ACE-2 therapy, immune modulators that prevent binding of virus to ACE 2 such as TACE inhibitors, TMPRSS2 inhibitors; IL-6 and TNF-α inhibitors warrants clinical trials as these modalities may be useful based on our epithelial-endothelial cross-talk model.