Predictors of COVID-19 Infection: A Prevalence Study of Hospitalized Patients

Abstract

Aim To find the predictors of coronavirus disease 2019 (COVID-19) in hospitalized patients. Methods A prevalence study compared the characteristics of COVID-19 patients with non-COVID-19 patients from January 19, 2020, to February 18, 2020, during the COVID-19 outbreak. Laboratory test results and pulmonary chest imaging of confirmed COVID-19 and non-COVID-19 patients were collected by retrieving medical records in our center. Results 96 COVID-19 patients and 122 non-COVID-19 patients were enrolled in this study. COVID-19 patients were older (53 vs. 39; P\u2009<\u20090.001) and had higher body mass index (BMI) than non-COVID-19 group (24.21\u2009±\u20093.51 vs. 23.00\u2009±\u20093.27, P\u2009=\u20090.011); however, differences in gender were not observed between the two groups. Logistic regression analysis showed that exposure history (OR: 23.34, P\u2009<\u20090.001), rhinorrhea (odds radio (OR): 0.12, P\u2009=\u20090.006), alanine aminotransferase (ALT) (OR: 1.03, P\u2009=\u20090.049), lactate dehydrogenase (LDH) (OR: 1.01, P\u2009=\u20090.020), lymphocyte (OR: 0.27, P\u2009=\u20090.007), and bilateral involvement on chest CT imaging (OR: 23.01, P\u2009<\u20090.001) were independent risk factors for COVID-19. Moreover, bilateral involvement on chest CT imaging (AUC\u2009=\u20090.904, P\u2009<\u20090.001) had significantly higher AUC than others in predicting COVID-19. Conclusions Exposure history, elevated ALT and LDH, absence of rhinorrhea, lymphopenia, and bilateral involvement on chest CT imaging provide robust evidence for the diagnosis of COVID-19, especially in resource-limited conditions where nucleic acid detection is not readily available.