Background: Cellular adhesion molecules (CAMs) foster interactions between endothelium and leukocytes. In HIV^- subjects, elevated soluble CAMs (sCAMs) predict cardiac events related to insulin resistance, diabetes, and dyslipidemia. However, in HIV⁺ patients, sCAMs rise as immune function worsens. High sCAMs also occur in HIV^- subjects with liver disease from chronic hepatitis C virus (HCV). As many HIV⁺ subjects have HCV co-infection and various cardiac risks, we studied the relative effects of immune, hepatic and metabolic function on sCAMs.

Methods: Fifty (50) HIV⁺ volunteers underwent fasting assay of sICAM-1, sVCAM-1, sE-, sL-, and sP-selectin, fasting lipid profile, CD4 count, HIV viral load (VL), ALT, AST, and HCV antibody. Each subject then completed an insulin suppression test to quantify insulin resistance, as well as a 75-g oral glucose tolerance, with assay of glucose, insulin at 0, 30, 60, 120,180 min for area-under-curve (AUC) calculations. All non-sCAM measures were used to stratify subjects into tertiles: sCAM levels were compared between those in the high tertile (n = 17) with those in the low tertile (n = 17) of each parameter.

Results: sICAM-1 was higher with high ALT and AST (both p < 0.02). sVCAM-1 was elevated with low CD4, high VL, high ALT and AST, and HCV⁺ (all p < 0.05). sE-selectin rose with high ALT, AST, and HCV⁺ (all p < 0.02). sL-selectin increased with high VL, ALT, AST and HCV⁺ (all p < 0.01). High sP-selectin was related to high VL (p < 0.05). sL-selectin and sVCAM-1 were higher with low total cholesterol (both p < 0.01). No other fasting or AUC metabolic parameter affected the sCAMs.

Conclusions: sICAM-1, sVCAM-1, sE- and sL-selectin in HIV infection are strongly related to chronic hepatitis, while sVCAM-1, sL-, and P-selectin are related to HIV viral burden and HIV-mediated immunodeficiency. Low cholesterol is associated with high sVCAM-1 and sL-selectin, likely due to advanced HIV and HCV co-infection, while no other metabolic factor impacts sCAMs. These results suggest that sCAMs do not usefully predict cardiac risk in HCV⁺ subjects.