Association between ABO blood groups and clinical outcome of coronavirus disease 2019: Evidence from two cohorts

Xianfei Zeng; Hongyan Fan; Dongxue Lu; Fang Huang; Xi Meng; Zhuo Li; Mei Tang; Jing Zhang; Nannan Liu; Zhixin Liu; Jingya Zhao; Wen Yin; Qunxing An; Xijing Zhang; Xingbin Hu

The Mann-Whitney U-test or the Kruskal-Wallis test, as appropriate, were used to compare the medians of continuous variables between or among groups. Categorical data were tested using a Given the limitations of a single factor analysis, we used multivariable logistic regression to further adjust the association of ABO blood type with occurrence of ARDS, AKI, or death for potential confounders, including the common risk factors of age, sex, pre-existing pulmonary disease, history of cardiovascular disease, diabetes mellitus, hypertension, and chronic kidney disease in the critical cohort. The history of diabetes mellitus was selected as a crucial confounder that needs to be adjusted because diabetes was reported to be protective against the development of ARDS, but as a risk factor for AKI and death.