LAUSR

Study Objectives: The primary objective is to evaluate the performance of pneumonia risk stratification tools in patients with COVID-19 for predicting adverse outcomes within 72 hours of admission The secondary objective is to evaluate pneumonia scores in predicting ICU admission, mortality and hospital revisits at 7 days An exploratory objective is to evaluate patient risk factors and pneumonia risk stratification scores on likelihood of intensive care unit (ICU) admission at 72 hours and 7 days Methods: This is a prospective observational study at an academic ED Adult patients who presented with dyspnea in the ED, were hospitalized with pneumonia symptoms and under investigation for COVID-19 were eligible for enrollment Demographic data, triage vital signs and laboratory data within 12 hours of presentation were collected via chart review and utilized for predictor variables Adverse outcomes were defined as patients needing intensive respiratory support or vasopressor support or death within 72 hours of admission MuLBSTA, CURB-65, and SMART-COP scores were evaluated using sensitivity, specificity and ROC curve analyses Regression analyses were performed to determine correlations between clinical predictors and ICU admission Results: Of 1196 patients, 139 tested positive for COVID-19 (11 6%) The COVID-19 positive cohort was comprised of 65 5% male with a mean age of 48 5 (±15) The most common comorbidities were hypertension (41 7%), smoking history (34 5%), and diabetes (31 7%) The most common chief complaints were cough (51 7%) and shortness of breath (50 4%), while common vital sign abnormalities were fever (24 5%), tachypnea (31 7%) and hypotension (8 6%) Adverse outcomes within 72 hours of arrival were noted in 33 (23 7%) patients with 31 (23%) admitted to the ICU Of these, 15 (10 7%) patients required non-invasive respiratory support while 21(15 1%) required endotracheal intubation Only 5 more patients (3 6%) required ICU care beyond 72 hours of arrival Overall mortality rate was 5% Of the three scores evaluated for adverse outcomes within the first 72 hours, CURB-65 was the only score with significant area under the curve (AUC 0 72 [95% CI 0 61,0 83], n=63), with sensitivity of 0 07 and specificity of 0 98 at a cut-off value of 3 (admission criteria per scoring system) For the 7-day outcome, once again, CURB-65 had a significant area under the curve (AUC 0 76 [95% CI 0 66, 0 86], p=0 00) with sensitivity of 0 06 and specificity of 0 98 with cut-off of 3, while SMART-COP and MuLBSTA did not reach significance (AUC 0 56 [95% CI 0 47, 0 71], p=0 15) and (AUC 0 55 [95% CI 0 43, 0 66], p=0 41) respectively Significant predictors for ICU admission include history of cancer, chronic kidney disease and diabetes A higher CURB-65 score was associated with increased odds of ICU admission at 72 hours and 7 days Conclusion: Current pneumonia risk stratification tools have poor sensitivities in detecting adverse outcomes in COVID-19 patients Hence, there is a need for new scoring tools to manage patients with COVID-19 related pneumonia in the ED