This study contrasts the sensitivity of four quantitative coronary angiography (QCA) measures (percent diameter stenosis [%S], minimum lumen diameter, average segment diameter, and percent involvement) in detecting 2-year treatment effects of two lipid-lowering therapies and reports on the longitudinal pattern after 4 years of treatment on the primary QCA trial endpoint (%S) for all, mild/moderate (<50%S), and severe lesions (> or =50%S). Patient cohorts were followed up from two randomized, placebo-controlled clinical trials of lipid-lowering therapies-colestipol/niacin in the Cholesterol Lowering Atherosclerosis Study (CLAS) and lovastatin in the Monitored Atherosclerosis Regression Study (MARS). Identical QCA methodology was used. In CLAS, the largest 2-year treatment effect size (=0.60) was noted for %S. In MARS, equivalent 2-year effect sizes (=0.15) were noted for three QCA measures. The largest 2-year effect size in %S was found in CLAS for mild/moderate lesions (=0.55) and in MARS for severe lesions (=0.31). Treatment in CLAS led to regression of disease in the first 2 years; treatment in MARS slowed progression of disease in the first 2 years and led to regression of disease after 4 years. Colestipol/niacin reduced progression of mild/moderate and severe lesions over the first 2 years of therapy; lovastatin reduced the progression of severe lesions over the last 2 years of therapy. We conclude that reducing the progression of atherosclerosis is not a simple proposition; maximal therapy for reducing and stabilizing atherosclerosis most likely will result from the selection of agents targeted at specific lesions.