Healthy postmenopausal women were randomly assigned to groups receiving 28-day treatment cycles of estradiol (E2) valerate (2 mg, days 1-21) combined with medroxyprogesterone acetate (10 mg, days 12-21) (N = 18), 17 beta-estradiol (1.5 mg, days 1-24) combined with desogestrel (150 micrograms, days 13-24) (N = 20), or placebo (N = 18). The progestational effects on the endometrium were assessed by histology, uterine bleeding pattern, and biochemical markers of secretion measured in endometrial tissue (E2 and isocitrate dehydrogenase) and serum (placental protein 14). After 2 years of therapy, 24 women in the hormone groups had secretory endometrium and 13 had atrophic endometrium; in the placebo group, the results were one and 15, respectively. Withdrawal bleeding generally started between days 9-12 after the addition of progestogen in the E2-medroxyprogesterone acetate group, and between days 14-17 in the E2-desogestrel group. All three biochemical markers of secretion were increased in each of the hormone-treated groups compared with the placebo group (P less than .01-.001). Serum placental protein 14 was twice as high in the secretory as in the atrophic phase (P less than .01). Isocitrate dehydrogenase, but not E2 dehydrogenase, was also higher in the secretory phase (P less than .05). Only serum placental protein 14 was significantly related to the uterine bleeding pattern (P less than .01). We conclude that serum placental protein 14 reflects both endometrial histology and bleeding pattern and may be a useful marker of progestational effects on the endometrium. The markers of secretion measured in endometrial tissue are not as reliable for endometrial histology or bleeding pattern.