The predictive values of dynamic blood lipid profile for mortality in COVID-19 patients

Abstract

Background Lipid metabolism is particularly affected in response to acute infectious diseases caused by viruses, bacteria, or parasites. The association between levels of lipid profiles and mortality in COVID-19 patients has become the subject of increasing interest. Objective To evaluate the predictive capacity of dyslipidemia for COVID-19 mortality based on dynamic data. Methods we conducted a retrospective, observational study, involving 135 COVID-19 patients admitted between January 1 and March 26, 2020. Results We found that non-survivals with COVID-19 displayed persistent dyslipidemia, including lower levels of total cholesterol (TC), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) with higher levels of triglycerides during the early stages of hospitalization. Notably, both the absolute values or the changes of TC, HDL, and LDL were closely related to mortality, and the AUCs of these three indicators at all time points and their changes were greater than 0.7. Notably, the values of AUCs of TC, HDL, and LDL at week 3 were 0.891, 0.895, and 0.879, while the AUCs for change of TC, HDL, and LDL were 0.975, 0.950 and 0.925 at week_3 − 1. Spearman correlation analyses showed that TC, HDL and LDL were significantly associated with CRP, D-dimer, BUN, CK and BNP at all four time points. Conclusion Blood lipid levels in the third week and changes from week 1 to week 3 are critical for predicting the mortality of COVID-19. Healthcare providers should pay close attention to the dynamic changes in lipid levels of COVID-19 patients.