Predictors of ICU Admission and Mortality in Patients with Coronavirus Disease - 2019 (COVID 19) in Community Hospitals

Abstract

Introduction: Coronavirus Disease 2019 (COVID-19) is caused by novel coronavirus Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). It was initially identified as the cause of pneumonia cases in Wuhan, China and has now rapidly spread throughout the world causing a pandemic. Although, 81% of patients have mild disease (pneumonia), 14% could have severe disease leading to hospitalization and 5% end up in intensive care unit. The mortality of patients in ICU is variable and has been reported to be as high as 80%, particularly the patient who require intubation. Not much is known about the factors leading to progression of hospitalized patient needing ICU care and the predictors of mortality among ICU patients. We did the univariate followed by multivariate logistic regression analysis to determine the predictors of mortality in ICU. Method: Retrospective data were collected from consecutive 101 patients admitted from March, 2020 to June, 2020. Data were collected from 5 different community hospitals in Eastern Virginia with varied demographics. Univariate and multivariate logistic regression was done to determine the factors associated with progression of hospitalized patient to ICU and the predictors of mortality in ICU. Result: Total 101 consecutive hospitalized patients in 5 community hospitals in Eastern Virginia were enrolled in the study. Total 52/101 patients were admitted into the ICU for respiratory failure. Of these, 40 patients required intubation and mechanical ventilation. Altogether, 32/52 patients died. Of these 32 patients, 25 had required intubation. Total 22/25 (88%) intubated patients passed away while 3 were successfully extubated. Of these 32 patients, one had mild ARDS, 6 had moderate ARDS and 18 had severe ARDS. Patients aged 60 years and above accounted for >2/3rd of the cases in ICU;mortality rate was higher in this age group as well. The inflammatory markers (CRP, D-dimer, Ferritin) peaked on day 8. The medications like Hydroxychloroquine, Azithromycin, Tocilizumab and Remdesivir did not alter the outcomes. Logistic regression analysis (univariate and multivariate) were done in the patients to determine the predictors of ICU admission from floor or ED. Logistic regression analysis was also done in the patients admitted to the ICU to look for the predictors of mortality. Conclusion: Based on logistic regression, none of the demographics (age, sex, race), symptoms, laboratory findings, chest imaging, ventilator settings or treatment identified the predictors of mortality in ICU in patients with COVID 19.