- In this retrospective study, 263 critically ill patients with COVID-19 acute respiratory distress syndrome (ARDS) were divided into two phenotypes based on 30 clinical variables
- The minority phenotype (Class 2, 27% of patients), was defined by increased markers of end-organ dysfunction and increased markers of vascular dysregulation suggestive of altered coagulation
- The odds of 28-day mortality in the class 2 phenotype were more than double that in the class 1 phenotype (40% vs 23%; OR = 2.2; 95% CI, 1.2–3.9)
- There was little distinction between phenotypes according to respiratory mechanics or severity of ARDS
- The hyperinflammatory phenotype established in previous studies among "classical" ARDS cohorts was not noted here
Prior investigations of acute respiratory distress syndrome (ARDS) divided patients into those with and without evidence of a hyperinflammatory response. In multiple datasets, the hyperinflammatory subgroup has consistently demonstrated 20% higher mortality.
Subscribe to the latest updates from Pulmonary & Critical Care Advances in Motion
Researchers at Massachusetts General Hospital investigated phenotypic groupings among patients with ARDS secondary to COVID-19. They found evidence for two distinct phenotypes—but the one associated with greater mortality is characterized by renal and cardiac impairment, altered coagulation and only mild relative hyperinflammation.
Leading author Sylvia Ranjeva, MD, PhD, physician in the Department of Anesthesia, Critical Care and Pain Medicine, Corey Hardin, MD, PhD, physician in the Division of Pulmonary and Critical Care Medicine at Massachusetts General Hospital, Taylor Thompson, MD, director of Critical Care Translational Research, Lorenzo Berra, MD, anesthesiologist in the Department of Anesthesia, Critical Care and Pain Medicine and medical director for Respiratory Care, and colleagues present their findings in EClinicalMedicine.
The researchers reviewed data on 263 patients with COVID-19 who were admitted to any Mass General ICU with acute respiratory failure between March 13 and August 2, 2020, were intubated within three days, and had minimum PaO2:FiO2 < 300 mmHg.
A multivariate mixture model was used to identify two distinct phenotypes based on 30 variables of interest, including demographic markers, body mass index, respiratory parameters, laboratory data and vasopressor requirement. 73% of patients were allocated to class 1 and 27% to class 2.
Three-class and four-class models showed no strong evidence of additional phenotypes.
Comparison of Phenotypes
Class 2 was distinguished from class 1 primarily by:
- Significantly increased markers of coagulopathy (e.g., elevated D-dimer, prothrombin time and activated partial thromboplastin time with decreased fibrinogen)
- Significant differences in markers of end-organ dysfunction (decreased serum pH and dramatically increased lactate, creatinine and troponin-T)
There was little distinction between phenotypes according to respiratory mechanics or ARDS severity (as judged from PaO2:FiO2).
Class 2 patients showed mild hyperinflammation compared with class 1 patients, but interleukin-6 levels were considerably lower than those reported for the hyperinflammatory phenotype in "classic" ARDS.
Interventions and Mortality
- The need for continuous renal replacement therapy was significantly more common in class 2 and corticosteroid use was significantly more common in class 1
- The two classes did not differ in the rate of use of antibiotics, rate of use of inhaled nitric oxide or average amount of heparin used
- Allocation to class 2 was associated with substantially increased odds of 28-day mortality (40% vs. 23%; OR, 2.2, 95% CI, 1.2–3.9); the difference was similar when the analysis was limited to patients who received corticosteroids
Toward the Future
These findings suggest phenotypic variation in COVID-19 ARDS is determined not just by respiratory parameters, but also by systemic and extrapulmonary processes, particularly altered coagulation. Larger, multicohort studies are planned with the goal of determining which variables differ most between the two phenotypes—a way to help frontline clinicians with prognostication.
Learn more about the Department of Anesthesia, Critical Care and Pain Medicine
Learn about research in the Division of Pulmonary and Critical Care Medicine