Abstract: Objective A prediction model for dyslipidemia was established based on laboratory data of patients with moderate coronavirus disease 2019 (COVID-19). Methods Clinical data and the results of the first laboratory examination of 74 patients with moderate COVID-19 at admission were collected. Patients were divided into normal and abnormal groups based on lipid levels. Univariate Logistic regression and colinear diagnosis were used to determine the risk factors of dyslipidemia, and backward stepwise regression was used to establish a line graph model. Receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis were used to evaluate the model discrimination, goodness of fit, and clinical usefulness. Results The patients with dyslipidemia at admission returned to normal levels of blood lipid at discharge. Logistic regression analysis showed that uric acid (OR=1.024) and alkaline phosphatase (OR=1.417) were combined to predict dyslipidemia (P<0.05). The ROC curve showed that the model had a high degree of differentiation (area under the curve was 0.987). The calibration curve showed that the model had a good fitting degree (P=0.989). The decision curve showed that the mode had a good clinical usefulness. Conclusion The combined model of uric acid and alkaline phosphatase established in this study can better predict dyslipidemia in patients with moderate COVID-19. This study is in the preliminary exploratory stage, and the clinical usefulness of the model needs further validation with large sample data, which can be used as a reference for related studies.

Key words: Corona virus disease 2019, Dyslipidemia, Uric acid, Alkaline phosphatase, Prediction model