Evaluation of colestipol/niacin therapy with computer-derived coronary end point measures. A comparison of different measures of treatment effect.

BACKGROUND: The Cholesterol Lowering Atherosclerosis Study has demonstrated beneficial effect of colestipol/niacin on coronary atherosclerosis using a panel-determined global coronary change score. We now report treatment group comparisons using quantitative coronary angiographic (QCA) measures from all processable segments in 85 of 162 randomly selected baseline/2-year film pairs. METHODS AND RESULTS: Treatment benefit was established for percent stenosis for either continuous or categorical analyses with regression established regardless of the per-patient scoring procedure. In addition, treatment benefit favoring regression was established in some cases for roughness and for percent involvement, a longitudinal estimate of the percent of coronary surface involved by raised lesions. Benefit on minimum diameter was directly related to whether the segment was proximal to a graft insertion and hemodynamically related to the bypass graft. QCA correlates of panel-determined progression were increases in percent stenosis and numbers of occluded lesions in native arteries and the number of progressing lesions in bypass grafts. CONCLUSIONS: These results demonstrate that a variety of computer measures can be used as end points in coronary angiographic therapy trials, but change in percent stenosis correlates best with visual panel assessments and best reflects the treatment benefit; when treatment effect sizes are moderate to large, the required sample size of coronary angiographic trials can be reduced when QCA is used.