Abstract
Background
The association between abacavir (ABC) and cardiovascular disease (CVD) risk in HIV-infected individuals is unclear. Putative mechanisms for an effect of ABC on CVD risk including endothelial dysfunction have been proposed; however, a biological mechanism has not been established.
Methods
This was a cross-sectional study of HIV-infected subjects with HIV RNA levels <400 copies/ml, who were randomly assigned to ABC or tenofovir (TDF) as initial therapy during a prior clinical trial. A small cohort of subjects on zidovudine (AZT; not randomly assigned) were studied to explore long-term exposure to this agent. All underwent brachial artery ultrasound for flow-mediated dilation (FMD), and D-dimer, high-sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6) and fasting lipids were measured. Between-arm differences were evaluated by multivariable linear or logistic regression modelling.
Results
There were 148 subjects (46 on ABC, 72 on TDF and 30 on AZT). Demographic characteristics were balanced across the groups except, as expected, AZT-treated participants were older, had higher CD4+ T-cell counts, and longer antiretroviral therapy duration. After adjusting for age, brachial artery diameter, and treatment duration, FMD was similar in those on ABC (3.9%) and TDF (5.4%; P=0.181). FMD was higher in those on AZT (6.1%; P<0.005). Levels of IL-6, hsCRP and detectable D-dimer were similar between groups.
Conclusions
Among individuals assigned to ABC or TDF in randomized clinical trials there were no significant differences in FMD or markers of inflammation and coagulation. Whether ABC contributes to risk of CVD remains unclear, but our results suggest that endothelial dysfunction, heightened inflammation, and altered coagulation are unlikely to be mechanisms by which the drug could increase CVD risk above that seen with TDF.