HCV genotype-related serum lipid differences and their adjunctive value for identifying HIV/HCV co-infection in a Chinese male cohort

Background Human immunodeficiency virus (HIV) and hepatitis C virus (HCV) infections are associated with altered lipid profiles. This study examined lipid differences across infection groups and their associations with HCV genotype. An exploratory analysis evaluated models for identifying HCV co-infection among people with HIV. Methods This retrospective study included 551 male participants: 119 healthy controls, 120 with HCV mono-infection, 156 with HIV mono-infection, and 156 with HIV/HCV co-infection. Generalized linear models compared seven lipid parameters and tested interactions between HCV genotype and infection status. Models were adjusted for age, body mass index, glucose, albumin, fibrosis-4 index (FIB-4), and estimated glomerular filtration rate. Multivariable linear regression assessed associations with HCV RNA levels. Three logistic regression models combined age with penalized regression-selected lipid parameters, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), or both. Model performance was assessed using receiver operating characteristic (ROC) analysis, calibration curves, and bootstrap internal validation. Results All seven lipid parameters were lower in the HIV/HCV co-infection group than in the HIV mono-infection group after adjustment (all P < 0.05). Adjusted mean total cholesterol (TC) was 3.70 (95% CI: 3.59-3.83) versus 4.53 (4.44-4.63) mmol/L, and low-density lipoprotein cholesterol (LDL-C) was 2.05 (1.97-2.14) versus 2.65 (2.57-2.73) mmol/L. Significant interactions between HCV genotype and infection status were observed for TC and LDL-C ( P = 0.006; P = 0.008). In the co-infection group, genotype 3 had lower TC and LDL-C than genotypes 1 and 2, with adjusted means of 3.11 and 1.58 mmol/L, respectively. Neither parameter differed significantly across genotypes in HCV mono-infection. HCV RNA levels were not significantly associated with any lipid parameter. In the exploratory prediction analysis, the combined model had an area under the ROC curve (AUC) of 0.977 (95% CI: 0.963-0.992), compared with 0.872 for the lipid model and 0.946 for the liver enzyme model. Its bootstrap-validated AUC was 0.975. Conclusion HIV/HCV co-infection was associated with marked reductions in serum lipid levels. Genotype-related differences in TC and LDL-C varied by HIV co-infection status. A model combining age, five lipid parameters, ALT, and AST may provide adjunctive information for identifying HCV co-infection among people with HIV.