To test our hypothesis that erythrocytes may induce plaque vulnerability and investigate the mechanism involved, we established a novel model of intraplaque hemorrhage in 56 New Zealand white rabbits with established plaques. Three distinct abdominal aortic plaques with similar thickness were identified in each rabbit with use of intravascular ultrasound (IVUS) imaging. Rabbits were equally divided into 4 groups depending on dosage of treatment; with the guidance of IVUS, one of the three plaques from each rabbit was injected from adventitia with autologous erythrocytes (RBC) or cholesterol (CH) for the following groups: RBC, 50 μL or 100 μL, and CH, 50 μL or 100 μL. One of the other two plaques in each rabbit received an equal volume of normal saline (NS) and one received no injection. Plaques in the 100 μL RBC group had a higher plaque rupture rate than its respective NS or blank controls plaques (57.1% vs. 14.3% or 14.3%, P<0.05). Plaques from the RBC or cholesterol groups showed, dose-dependently, more macrophage infiltration, more superoxide and lipid content, thinner plaque fibrous cap, higher mRNA level of MCP-1, IL-1 or IFN-gamma and higher vulnerability index than controls, especially in the RBC group. Thus, erythrocyte treatment can dose-dependently induce the vulnerability of plaques. Accumulation of lipid content and augmentation of oxidative stress and inflammation in the plaques are the probable pathological mechanisms involved.