ABSTRACT
Objective This study aimed to determine the association of hyperlipidemia with clinical endpoints among hospitalized patients with COVID-19, especially those with pre-existing cardiovascular diseases (CVDs) and diabetes. Methods This multicenter retrospective cohort study included all patients who were hospitalized due to COVID-19 from 21 hospitals in Hubei province, China between December 31, 2019 and April 21, 2020. Patients who were aged < 18 or ≥ 85 years old, in pregnancy, with acute lethal organ injury (e.g., acute myocardial infarction, severe acute pancreatitis, acute stroke), hypothyroidism, malignant diseases, severe malnutrition, and those with normal lipid profile under lipid-lowering medicines (e.g., statin, niacin, fenofibrate, gemfibrozil, and ezetimibe) were excluded. Propensity score matching (PSM) analysis at 1:1 ratio was performed to minimize baseline differences between patient groups of hyperlipidemia and non-hyperlipidemia. PSM analyses with the same strategies were further conducted for the parameters of hyperlipidemia in patients with increased triglyceride (TG), increased low-density lipoprotein cholesterol (LDL-C), and decreased high-density lipoprotein cholesterol (HDL-C). Mixed-effect Cox model analysis was performed to investigate the associations of the 28-days all-cause deaths of COVID-19 patients with hyperlipidemia and the abnormalities of lipid parameters. The results were verified in male, female patients, and in patients with pre-existing CVDs and type 2 diabetes. Results Of 10 945 inpatients confirmed as COVID-19, there were 9822 inpatients included in the study, comprising 3513 (35.8%) cases without hyperlipidemia and 6309 (64.2%) cases with hyperlipidemia. Based on a mixed-effect Cox model after PSM at 1:1 ratio, hyperlipidemia was not associated with increased or decreased 28-day all-cause death [adjusted hazard ratio (