To examine the effects of estrogen replacement on lipids and angiographically defined coronary artery disease (CAD) in postmenopausal women, lipid profiles were obtained in 90 consecutive postmenopausal women undergoing diagnostic coronary angiography. Eighteen women (20%) were receiving estrogen and 72 (80%) were not. CAD (defined as greater than or equal to 25% luminal diameter narrowing in a major coronary artery) was present in only 22% of women (4 of 18) receiving estrogen and in 68% (49 of 72) who were not (p less than 0.001), with an odds ratio of 0.13. Mean high-density lipoprotein (HDL) cholesterol level was significantly higher (63 +/- 6 vs 48 +/- 2; p less than 0.01) and mean total/HDL cholesterol ratio significantly lower in women receiving estrogen than in those who were not (4.2 +/- 0.5 vs 5.1 +/- 0.2; p less than 0.05). The other lipid values were similar in both groups. On multiple logistic regression analysis, absence of estrogen use was the most powerful independent predictor of the presence of CAD (p less than 0.001), with total/HDL cholesterol ratio as the only other variable selected (p less than 0.01). Thus, among 90 consecutive postmenopausal women undergoing diagnostic coronary angiography, estrogen replacement therapy was associated with an 87% reduction in the prevalence of CAD, and those receiving estrogen had a significantly higher mean HDL cholesterol level and lower mean total/HDL cholesterol ratio.