Implications of Lung Compliance Phenotypes in COVID-19-Induced Acute Respiratory Distress Syndrome

Background: Early literature suggested a phenomenon where coronavirus disease 19 (COVID-19)-induced acute respiratory distress syndrome (ARDS) presented with high respiratory system compliance (RC) in some patients. Thus, it posed a challenge to conventional ARDS ventilator management strategies. Research question: Are there different phenotypes of lung compliance among mechanically ventilated patients with COVID-19-induced ARDS? Does the RC phenotype impact the clinical outcome? Study design and methods: This was a single-center historical cohort study that included patients who were mechanically ventilated for COVID-19-induced ARDS and who met the Berlin Criteria for ARDS. RC was calculated at intubation, for eight weeks, and at extubation. Data on demographic factors and comorbidities were also collected. Measurements: RC was calculated by tidal volume/pressure. Descriptive statistics, chi-squared test, Student’s t-test, and multivariable logistic regression were used to analyze the data. Main results: Compliance groups were derived from the collected data. Most patients (73.6% to 89.4%) had severely reduced (≤40 ml/cmH₂O) static RC, and (10.7% to 26.4%) had RC ≥40 ml/cmH₂O. At week one and at extubation, there was a statistically significant difference in mortality between high versus moderate versus low RC groups: week one: (57.1% versus 66.7% versus 87.7%, respectively, p = 0.02); extubation: (83.3% versus 58.8% versus 87.2%, p = 0.01, respectively). Remdesivir administration, older age, and higher Charlson Comorbidity index scores were associated with increased mortality (4.2 OR 95% CI [1.61, 10.87]), (1.03 OR 95% CI [1.003, 1.07]), and (1.31 OR 95% CI [1.03, 1.67]), respectively. Declining PaO2/FiO2 (P/F) ratios were associated with lower RC. Interpretation: COVID-19 ARDS presents a range of RC and different phenotypes, which may require different ventilator strategies. Given the increased mortality at extubation in patients with low RC, it poses the possibility of a bedside prognostication tool. Older age, number of comorbidities, and Remdesivir use were associated with increased mortality. Lastly, declining P/F ratios were associated with lower RC.