Host susceptibility to severe COVID-19: a retrospective analysis of 487 case outside Wuhan

Yu Shi,Xia Yu,Hong Zhao,Hao Wang,Ruihong Zhao,Qun Cai,Shanshan Sun,Hong Fang,Yijie Wang,Zhangmin Hu,Jinnan Duan,Xiaoman Lin,Huilan Tu,Sheng Tu,jifang sheng
DOI: https://doi.org/10.21203/rs.3.rs-16021/v1
2020-01-01
Abstract:Abstract Background: The recent outbreak of SARS-CoV-2 infection results in a considerable morbidity and mortality, mainly in China. The study is to investigate the intrinstic features of infected patients that associated with severe type of this disease. Method: A total of 487 laboratory-confirmed COVID-19 patients were included in analysis. The demographic and epidemiological of patients representing as mild and severe at admission were compared. A step-wise multivariate logistic regression analysis were performed to identify significant risk factors associated with severe COVID-19 . A score systemc incorporating risk factors was established for risk stratification and validated in a small cohort during in-hospital follow-up. Results: Of all patients, 49 (10.1%) cases are severe at admission. Severe cases are elder [56(17) vs. 45(19), P<0.001), with more male (73.5% vs. 50.9%, P=0.003). They have a higher incidence of hypertension (53.1% vs. 16.7%, P<0.001), diabetes (14.3% vs. 5.0%, P=0.009), cardiovascular diseases (8.2% vs. 1.6%, P=0.003) and malignancy (4.1% vs. 0.7%, P=0.025), and less exposure to epidemic area (49.0% vs. 65.1%, P=0.027), but more infected familiy members(P=0.031). On multivariate analysis, elder age, male and presence of hypertension are independently associated with severe disease at admission. A host risk score, incorporating age, sex and hypertension history, clearly stratifies risk of developing severe type of COVID-19 both in patients at admission and during in-hospital follow-up. Conclusions: Elder age, male and presence of hypertension are associated with host susceptibility to developing severe COVID-19. The host risk score may be a useful tool to identify high risk indiviuals but requires validation.
What problem does this paper attempt to address?