ABSTRACT
Introduction: There has been substantial lay press interest in whether patients who die of COVID-19 are dying of COVID-19 or other underlying illnesses. We have previously described a clinician scoring system that delineates whether a patient's prognosis is driven by acute illness vs comorbidities. In this study we compared this clinician phenotyping system in patients with COVID-19 to ICU biobank patients enrolled prior to 2020. Methods: Clinician phenotyping was performed on patients enrolled into 2 Stanford ICU biobanks: (1) Patients admitted to Stanford ICU before 2020 with at least one ARDS risk factor (non-COVID-19 cohort) and (2) Patients admitted to the ICU with confirmed COVID-19 infection (COVID-19 cohort). The roles of each of the following categories in determining prognosis at time of enrollment were scored from 1-5 (1 being non-contributory and 5 being highly contributory): goals of care, comorbidities, baseline functional status, acute multi-system organ failure (MSOF), ARDS, acute neurologic injury, and terminal illness. Each patient's primary determinant of mortality was then grouped into one of four categories: (1) MSOF or ARDS (>=4 for ARDS or multi-organ failure), (2) Comorbidities (>=4 for goals of care, comorbidities, acute neurologic injury, or baseline functional status), (3) Mixed (>= 4 for both of the above), or (4) None (<4 for all categories). The odds ratio for prognostic subgroupings was compared between patient cohorts.Results: Clinician phenotyping was performed on 70 non-COVID-19 patients admitted to the ICU with at least one risk factor for ARDS, and 28 patients with COVID-19 admitted to the ICU. Compared to the general ICU population, COVID-19 patients were younger (median age 53 vs 71), more likely to be male (71% vs 54%), and more likely to identify as Hispanic (64% vs 13%). 60-day mortality was higher in the COVID-19 group (29% vs 20%). ARDS or MSOF was the primary determinant of prognosis in 64% of patients with COVID-19 compared to 31% of traditional ICU patients (OR 3.9, 95% CI 1.6-9.9, Figure 1). Only 11% of COVID-19 patients' primary determinant of prognosis was comorbidities or goals of care relative to 31% in the traditional ICU population (OR 0.26, 95% CI 0.07-0.9).Conclusion: In this study, ARDS or MSOF is far more likely to be the main risk factor for death in patients with COVID-19 than in a traditional ICU population. This suggests that most COVID-19 fatalities are attributable to acute infection with SARS-CoV-2 rather than progression of underlying disease.
ABSTRACT
Introduction: COVID-19 respiratory infections are associated with copious, adherent respiratory secretions that prolong chronic ventilation and contribute to the morbidity and mortality caused by the disease. To target thinning these secretions requires understanding the physical properties that contribute to these tenacious secretions. Methods: We received surrogate consent from 11 patients with COVID-19 ARDS to collect an undiluted tracheal aspirate obtained as part of routine care. All samples were frozen at-80C immediately after collection and heat inactivated at 65C for 30 minutes to render them noninfectious prior to further analyses. These were compared with sputum collected from 4 patients with Cystic Fibrosis and 15 healthy controls. Clinical phenotypes of COVID patients included ARDS severity, treatments related to sputum clearance, and patient outcomes including length of mechanical ventilation and mortality. Sputum was characterized for percent solids, double strand DNA, and hyaluronin content. Results: Sputum samples were collected from 11 patients intubated with COVID-19 ARDS, with aspirates sampled between 2-18 days into their course. 63% of patients had thick secretions described by respiratory therapy, 81% had moderate or severe ARDS, and 90% had prolonged mechanical ventilation > 14 days. We found that the physical characteristics of COVID-19 sputum are similar to CF sputum, with markedly higher percent solid, hyaluronin, and double strand DNA content than that seen in sputum from healthy controls (Figure 1). Conclusion: In this proof of concept study, we show the feasibility of measuring numerous physical characteristics of COVID-19 sputum. As expected, we found that percent solids, DNA content and hyaluronin were similar to the known thick secretions of CF, and markedly increased in comparison to sputum of healthy controls. We were unable to identify differences in outcomes in this small cohort, with collection of additional samples ongoing to improve power. Figure 1. Physical characteristics of COVID sputum are much more similar to CF than normal sputum.