Relationship of estradiol levels to breakthrough bleeding during continuous combined hormone replacement therapy

Publisher Website
Google Scholar
Abstract
Objective: To determine whether serum estradiol and dydrogesterone concentrations are associated with the occurrence of breakthrough bleeding. Methods: In a prospective, double-blind study, 194 postmenopausal women were allocated randomly to receive one of four doses of dydrogesterone (2.5 mg, 5 mg, 10 mg, 15 mg) continuously combined with 2 mg of micronized 17β-estradiol. All medication was taken orally for a total of 168 days. Vaginal bleeding was recorded on a daily basis. Serum estradiol (E2) and dihydrodydrogesterone (the main metabolite of dydrogesterone) trough levels were measured at day 85 and at the end of the study (day 168). Bleeding pattern analysis was done according to the reference period method. Results: One hundred fifty-two of 177 women who completed the study supplied valid data on drug compliance, smoking habits, bleeding episodes, and serum hormone concentrations, which were used to assess the impact of serum E2 and dihydrodydrogesterone concentrations on the occurrence of breakthrough bleeding. Logistic regression analysis identified only the serum E2 concentration as having an independent, statistically significant effect (P = .003) on the occurrence of breakthrough bleeding; no such effect was associated with dihydrodydrogesterone levels (P = .118). The relative risk for the occurrence of breakthrough bleeding was 2.7 (95% confidence interval [CI] 1.454, 5.609) for serum E2 concentrations greater than 40 pg/mL. Conclusion: The occurrence of breakthrough bleeding during continuous combined hormone replacement therapy with estradiol and dydrogesterone in postmenopausal women was related to serum estradiol levels and not to dydrogesterone levels. Further studies are needed to test the hypothesis that estrogen is a major factor in the incidence of bleeding during postmenopausal hormone replacement therapy.

This publication has 0 references indexed in Scilit: