ABSTRACT
Rationale: A dysregulated host inflammatory response in COVID-19 is considered a central pathogenetic mechanism of acute lung injury and extrapulmonary end-organ damage. However, limited comparative data are available as to whether the host-response in COVID-19 ARDS differs from patients with other (non-COVID) ARDS etiologies, and how such differences may inform targeted immunomodulating therapeutics. Methods: We prospectively enrolled 36 intubated patients with COVID-19 ARDS and 70 hospitalized non-intubated patients with COVID-19 (COVID-19 non-ARDS), and compared them with a pre-COVID-19 cohort of patients with bacterial (n=21), viral (n=14), and culture-negative ARDS (n=30). We measured 10 host-response biomarkers of innate immunity and epithelial/endothelial injury (IL-6, IL-8, IL-10, RAGE, TNFR1, Angiopoeitin-2, Procalcitonin, Fractalkine, Pentraxin-3, ST2) in plasma. Using a 4-variable predictive model (TNFR1, Angiopoeitin-2, Procalcitonin and bicarbonate levels), we classified patients into hyper-vs. hypo-inflammatory subphenotypes. We compared biomarker levels, subphenotypes and outcomes between the clinical groups. Results: Host-response biomarker levels were widely distributed between the 5 groups, with a characteristic pattern for IL-6, IL-8, Angiopoeitin-2, Procalcitonin, ST-2 and fractalkine: COVID-19 ARDS patients had higher biomarker levels than COVID-19 non-ARDS (p<0.01), lower levels than bacterial or culture-negative ARDS (p<0.01), and similar levels to viral ARDS (Figure 1A example for IL-6). A lower proportion of the COVID-19 ARDS cohort was classified in the adverse hyper-inflammatory subphenotype (15%) compared to bacterial (47%) and culture-negative ARDS (31%) (Figure 1B). Despite the lower level of inflammatory host responses, COVID-ARDS patients had longer median duration of mechanical ventilation (20.5 [10.0-40.8] days) compared to bacterial (8.0 [5.0-25.0]), culture-negative (7.0 [5.2-9.8]) and viral ARDS (7.5 [3.5-14.8]) (p<0.01). Patients with COVID-19 but without ARDS had lower 30-day mortality (6%) compared to patients with ARDS from COVID-19 (31%) or other etiologies (bacterial 33%, culture-negative 40% and viral 21%). Conclusion: Development of ARDS from COVID-19 is characterized by intensified inflammation compared to hospitalized COVID-19 patients not requiring mechanical ventilation. Compared to ARDS from other etiologies, host-response inflammatory profiles in COVID-19 ARDS appear similar to other viral etiologies of ARDS, and are lower compared to bacterial or culture-negative ARDS. The etiology of worse clinical outcomes of COVID-19 ARDS despite the lower frequency of the prognostically adverse hyper-inflammatory subphenotype warrant urgent investigation. (Table Presented).