Recent studies have identified the presence of estrogen receptor (ER)-beta in addition to ER-alpha in human breast cancers, but the clinicopathological characteristics of ER-beta -positive tumors remain to be established. In this study, we have conducted an immunohistochemical analysis of ER-alpha and ER-beta expression in human breast cancers. In addition, we investigated the correlation of ER-alpha and ER-beta expression with progesterone receptor (PR) status, determined by enzyme immunoassay, and with various clinicopathological factors. Of 79 tumors, 49 (62%) were positive for ER-alpha and 24 (30%) were positive for ER-beta, and there was no significant association between ER-alpha and ER-beta expression. ER-alpha -positive tumors were significantly more likely to be PR-positive than were ER-alpha -negative tumors (P.<0.0001), but there was no significant association between ER-<beta> expression and PR status. However, the PR values of ER-alpha -positive and ER-beta -positive tumors (65 +/- 17 fmol/mg protein, mean +/- SE) were marginally significantly lower than those of ER-alpha -positive and ER-beta -negative tumors (340 +/- 109) (P=0.08). ER-beta positivity was significantly associated with small tumor size (less than or equal to2 cm) and high histological grade (P<0.05), and this association was also observed when only ER-<alpha>-positive tumors were considered. These results suggest that determination of ER-beta status might be clinically useful for further defining the characteristics of ER-alpha -positive tumors.