Abstract: Objective To analyze the factors related to the occurrence of dyslipidemia in HIV-infected patients after antiviral therapy, and to establish a predictive model for the occurrence of dyslipidemia. Method Data of all HIV-infected patients from May 2016 to May 2021 at Beijing You'an Hospital Affiliated to Capital Medical University were selected. A total of 1000 samples were included as the modeling group, and divided into the dyslipidemia group (case group) 388 cases and the normol lipidemia group (control group) 612 cases, and including 125 HIV infected individuals treated at our hospital from June to December 2022 as the model validation group. Clinical indicators and laboratory examination data were collected, and the related factors of dyslipidemia were analyzed by binary Logistic regression, and a prediction model of risk factors was established. Result Among 1000 HIV-infected patients, 982 were males and 18 were females. The mean age was (36.6± 11.2) years. Binary logistic regression analysis showed age ≥ 50 years (OR=2.029, 95%CI: 1.966-2.074, P<0.001), body mass index (BMI) ≥ 24 kg/m² (OR=5.008, 95%CI: 1.216-5.429, P=0.035), non-integrase drug regimens for antiviral therapy (OR=1.126, 95%CI: 1.083-1.147, P=0.025), duration of antiviral treatment ≥ 3 years (OR=1.722, 95%CI: 1.203-1.945, P<0.001), fasting plasma glucose (FPG) ≥ 6.0 mmol/L (OR=1.303, 95%CI: 1.006-1.586, P=0.038) were risk factors for the occurrence of dyslipidemia in HIV-infected patients treated with antiviral therapy, and weekly exercise time ≥ 1 day (OR=0.487, 95%CI: 0.374-0.636, P<0.001) was an independent protective factor for the occurrence of dyslipidemia. There is no statistically significant difference in the scores of various indicators between the validation group and the modeling group in the risk prediction model, indicating that the model has a good fit. The area under the ROC curve of the risk prediction model for the modeling group was 0.918 (95%CI: 0.898-0.937, P<0.001). The cut-off value was 0.938, with a sensitivity of 85.60%, a specificity of 93.22%, a positive predictive rate of 89.90%, and a negative predictive rate of 91.51% for the occurrence of dyslipidemia. Conclusion The incidence of dyslipidemia in HIV-infected patients receiving antiretroviral therapy is 38.80%. Age, BMI, antiviral drugs, duration of antiviral therapy, weekly exercise time and FPG are the relevant influencing factors for the occurrence of dyslipidemia. The prediction model of risk factors based on independent correlation factors has good predictive value for the occurrence of dyslipidemia in HIV-infected patients.

Key words: Antiviral therapy, Human immunodeficiency virus-infected people, Abnormal blood lipids, Relevant factors, Prediction model