Studies of the effects of menopausal hormone therapy on coronary artery disease (CAD) in postmenopausal women have provided contradictory results. Although recent experimental studies have revealed no beneficial effect of combination therapy with estrogen (E) and progesterone (P), the effect of monotherapy with E remains unknown. We retrospectively examined the medical records of 843 consecutive women aged > or =55 years who underwent their first cardiac catheterization between January 1996 and December 1998. We compared the presence and severity of CAD, defined as > or =1 diseased coronary vessel (with stenosis > or =50%), in women who only took E, E+P, or no menopausal hormone therapy. In all, 210 women (33%) took hormones, of whom 47 (22%) used E+P and 163 (78%) used E only. Women who used any hormones tended to be healthier than nonusers, but E+P users had a lower prevalence of risk factors and co-morbidities than E users. In unadjusted analyses, both the E and E+P groups were significantly less likely to have CAD than nonusers (relative risk [RR] 0.71, 95% confidence interval [CI] 0.58 to 0.84 for the E group; RR 0.76, 95% CI 0.54 to 0.99 for the E+P group). Demographic factors, CAD risk factors, co-morbidities, and primary prevention medication use explained the association between E+P and the presence of CAD (RR 1.14, 95% CI 0.74 to 1.39). In contrast, adjustment for these factors had a minimal effect on the association between E and CAD compared with nonusers (RR 0.79, 95% CI 0.59 to 0.98). Thus, the apparent protective effect of combination menopausal hormone therapy with E+P on CAD is due to differences in other patient characteristics. In contrast, unopposed E therapy may have a protective effect on CAD.