Investigating Strategies for Mechanical Ventilation in COVID-19 via Computational Simulation of Virtual Patients

Initial clinical guidelines for mechanical ventilation of COVID-19 patients suggested following standard approaches used in the treatment of acute respiratory distress syndrome (ARDS). However, emerging clinical experience suggests that patients with COVID-19 pneumonia may present an atypical form of the syndrome, characterised (at least initially) by severe hypoxemia with relatively well-preserved lung mechanics (i.e. preserved lung compliance) - a combination that is rarely seen in ARDS [1]. The pathophysiological basis for this disease phenotype is currently unclear. A recent study also noted a significant time-related disease spectrum in COVID-19 patients, with at least two potential "sub-phenotypes": Type L, characterized by low elastance (i.e. high compliance), low ventilation to perfusion ratio, low lung weight and low recruitability by imaging; and a Type H, characterized by high elastance, high right-to-left shunt, high lung weight and high recruitability [2]. It is currently unclear whether a transition from Type L to Type H in some patients is due primarily to the evolution of the COVID-19 pneumonitis or to lung injury caused by injurious mechanical ventilation during the Type L phase.
These issues (and others) will require extensive investigation using experimental (animal) models and ultimately clinical trials in human patients. However, clinicians need insights into the possible underlying disease pathophysiology of COVID-19 now, in order to develop appropriate strategies for ventilating patients' lungs. We propose adapting a state-of-the-art computational simulator, which has been developed to investigate mechanical ventilation in ARDS [3-4], to investigate a range of issues that are specific to mechanical ventilation of COVID-19 patients.