Apolipoprotein B has been shown to be a better predictor of coronary heart disease than low-density lipoprotein (LDL) cholesterol, and non-high-density lipoprotein (non-HDL) cholesterol may also be a better parameter for coronary heart disease risk assessment and as a target for therapy. Data from the Atorvastatin Comparative Cholesterol Efficacy and Safety Study (ACCESS) were used to assess the correlation between lipid and apolipoprotein B levels before and after lipid-lowering therapy and to examine the effects of 5 hydroxymethylglutaryl coenzyme A reductase inhibitors on lipids and apolipoprotein B. The 54-week study randomized 3,916 hypercholesterolemic patients to atorvastatin, fluvastatin, lovastatin, pravastatin, or simvastatin, initiated at recommended starting doses with titrations as needed at weeks 6, 12, and 18 to achieve National Cholesterol Education Program LDL targets. Compared with LDL cholesterol, non-HDL cholesterol correlated better with apolipoprotein B levels at baseline (r = 0.914, p <0.0001) and at week 54 (r = 0.938, p <0.0001), and the correlation was strong across all baseline triglyceride strata. At starting doses, atorvastatin (10 mg) lowered non-HDL cholesterol by 33.3% compared with 26.6% with simvastatin (10 mg), 24.1% with lovastatin (20 mg), 17.2% with fluvastatin (20 mg), and 17.0% with pravastatin (10 mg). Atorvastatin also provided greater reductions in non-HDL cholesterol after dose titration, and a greater percentage of patients taking atorvastatin achieved non-HDL cholesterol targets. Baseline triglyceride did not affect non-HDL cholesterol reductions with any of the 5 hydroxymethylglutaryl coenzyme A reductase inhibitors. Fewer patients achieved non-HDL cholesterol targets than LDL cholesterol targets, particularly among high-risk patients, implying that if non-HDL cholesterol was used as a target for treatment, more patients would need to be treated more aggressively than National Cholesterol Education Program guidelines require.