BACKGROUND: Lowering low-density lipoprotein cholesterol (LDL-C) with statins reduces atherosclerosis. LDL and high-density lipoprotein (HDL) are commonly measured by their cholesterol content, but non-HDL cholesterol, LDL particle number (LDL-P), or total apolipoprotein B (apoB) may better predict cardiovascular risk. Few studies have examined relations among lipoprotein levels and composition before and after interventions to lower LDL-C and non-HDL-C. OBJECTIVE: To measure changes in carotid artery intimal media thickness (CIMT) and lipid concentration and composition during 36 months of statin therapy. METHODS: Analyses were conducted on 418 diabetic individuals, with complete data and no prior cardiovascular events, who were randomized to aggressive (AG) versus standard (STD) treatment for LDL-C, non-HDL-C, and systolic blood pressure (SBP) as part of the Stop Atherosclerosis in Native Diabetics Study (SANDS). RESULTS: The AG group achieved average LDL-C and non-HDL-C of 71mg/dL and 100mg/dL and a decrease in CIMT. No significant interactions were observed between treatment effect and initial levels of LDL-C, non-HDL-C, HDL-C, triglycerides, apoB, or LDL-P. Decreases in LDL-C (p<.005) and non-HDL-C (p<.001) were independently correlated with CIMT regression in the AG group. Changes in apoB and LDL-P showed borderline correlations with CIMT regression (p=.07 and p=.09). CONCLUSIONS: In diabetic adults with no prior cardiovascular events, treatment to current targets for lipids and SBP reduces atherosclerosis progression and when more aggressive targets are met, atherosclerosis regresses. The aggressive targets for LDL-C and non-HDL-C appeared to be the main determinants of CIMT regression and were more predictive of this outcome than changes in LDL-P or apoB.