Is Estriol Safe?

By Anthony R. Scialli, MD, and Adriane Fugh-Berman, MD

In the wake of the Women’s Health Initiative (WHI), practitioners can expect an upsurge in questions about safer alternatives to conjugated estrogens with progestin. Estriol, either on its own or as part of a mixed estrogen formulation, has long been promoted by some alternative medicine practitioners as a natural, safe form of estrogen therapy. Is there any evidence that estriol is safer than conjugated estrogens?

Estriol is an endogenous ovarian hormone available in pharmaceutical preparations to treat hot flashes or vaginal dryness.1 A weak, short-acting estrogen, more popular in Europe, estriol is not commonly used by conventional medical practitioners in the United States.

Tri-estrogen preparations (also called Tri-est) are mixtures of estrone, estradiol, and estriol; typically, the proportions are 80% estriol, 10% estradiol, and 10% estrone. Tri-est usually is administered in doses of 2.5-5 mg/d, either continuously or 25 days a month. No clinical trials have been performed on this preparation. Natural progesterone often is prescribed with this regimen (topical forms, however, do not achieve serum levels high enough to ensure endometrial protection, and should not be used for that purpose (see Alternative Therapies in Women’s Health, June 1999).

Estrogenic Effects on Endometrium

Estriol and Tri-est proponents claim that estriol does not cause endometrial stimulation, but evidence does not support this claim. A recent study claiming that estriol does not cause endometrial stimulation distorts its own data. A cross-sectional multicenter study of 241 postmenopausal women in Sweden compared 125 women who had taken oral estriol (1-2 mg) with 116 women who had not taken any hormone replacement therapy (HRT) for at least one year.2 All women underwent transvaginal ultrasound and endometrial biopsy. Mean treatment duration was 4.3 years (median 3 years). Fourteen women in the control group also had used estriol in the past for a period ranging from 0.5-15 years, but all had stopped at least a year prior to study entry.

Although the authors concluded that "No clinically relevant difference was found between the endometrium status [assessed by histology and TVS] and untreated controls. This trial supports the endometrial safety of maintenance treatment with oral estriol ¼ ," we disagree, based on the data presented in the study. The authors admit that mean endometrial thickness was significantly higher in the estriol group, but neglect to mention that the number of women with clinically important thickening (defined by them as > 4 mm by transvaginal ultrasound) also was increased in the estriol group. Histological diagnosis was obtained for 201 women. One endometrial cancer was found in the control group. Mean endometrial thickness was 3.7 mm in the estriol group (SD 2.83, range 1-19 mm), for controls 2.5 mm (SD 1.75, range 1-11 mm). Eighty-nine percent in the estriol group had an endometrium measuring £ 4 mm, compared to 97% of controls. Using the Fisher exact test, we got a P value of 0.0273, a significant difference between groups. Fourteen of 99 (14.1%) estriol users and three of 102 (2.9%) controls had polyps. Again, statistical comparison does not appear in the paper, but using the Fisher exact test results in a P value of 0.0048, providing additional support for our conclusion that endometrial stimulation is associated with estriol. The authors dismissed their own findings with the statement, "the histopathological diagnosis between the groups are not statistically significant."

As the risk of endometrial hyperplasia and cancer increase with duration of estrogen use, data should have been broken out by how long women had taken estriol. Providing only mean and median duration of use is not acceptable. Additionally, the cross-sectional design of the study would be expected to underestimate risk because it would not identify women who had discontinued estriol due to postmenopausal bleeding, endometrial hyperplasia, or other estrogen-related problems.

Oddly, a relevant previous study done by the same author is not mentioned in this recent paper. This study, of 1,110 women with irregular bleeding on HRT or postmenopausal bleeding while not on HRT, found that endometrial hyperplasia was significantly more common in women taking estriol than in women taking sequential estrogen and progestin therapy or women not receiving HRT.3

Unopposed estrogen stimulation of the endometrium is a risk factor for endometrial cancer, with which estriol also has been associated. A case-control study in Sweden compared 789 postmenopausal women with endometrial cancer with 3,368 controls. Five years of oral estriol 1-2 mg/d was associated with tripling the risk of endometrial cancer; the risk of atypical endometrial hyperplasia was increased 8.3-fold, compared to never-users.4 Ever-use of oral estriol was associated with a doubling of endometrial cancer risk (vaginal estriol was not associated with significantly increased in risk). An earlier prospective cohort study in Sweden found no increase in endometrial cancer rates among estriol users, but data on duration and recency of use were not available.5

Breast Cancer

Since the WHI found an increase in breast cancer risk with conjugated estrogens and medroxyprogesterone acetate, natural hormone proponents can be expected to resurrect the claim that estriol and Tri-est decrease breast cancer risk. These claims are based on papers published by Henry M. Lemon during the 1960s and 1970s that claimed that estriol had potential in preventing and treating breast cancer. A 1978 commentary by Follingstad6 cites an unpublished study by Lemon in which an unspecified number of postmenopausal breast cancer patients received between 2.5 and 15 mg of estriol for an unspecified amount of time; 37% were said to have had remission or arrest of metastasis. Lemon never published a paper with those figures in it, but in 1980 published a review on estriol in which he described giving oral estriol 5-15 mg/d to 24 subjects with breast cancer. No therapeutic effect was claimed. Despite the fact that six subjects (25%) experienced increased growth of metastases and two developed endometrial hyperplasia, the author’s enthusiasm for estriol appears to have remained undimmed.


There is no reasonable scientific evidence that estriol has anticancer effects or that it is safer than estradiol or conjugated estrogens. The most recent study adds to the growing body of data that show that unopposed oral estriol significantly increases the risk of endometrial pathology. Physicians should discourage the use of unopposed estriol or Tri-est. 


1. Cardozo L, et al. Meta-analysis of estrogen therapy in the management of urogenital atrophy in postmenopausal women: Second report of the Hormones and Urogenital Therapy Committee. Obstet Gynecol 1998;92(4 Pt 2):722-727.

2. Granberg S, et al. The effects of oral estriol on the endometrium in postmenopausal women. Maturitas 2002;42:149-156.

3. Granberg S, et al. Endometrial sonographic and histologic findings in women with and without hormonal replacement therapy suffering from postmenopausal bleeding. Maturitas 1997;27:35-40.

4. Weiderpass E, et al. Low-potency oestrogen and risk of endometrial cancer: A case-control study. Lancet 1999; 353:1824-1828.

5. Schairer C, et al. Cause-specific mortality in women receiving hormone replacement therapy. Epidemiology 1997;8:59-65.

6. Follingstad AH. Estriol, the forgotten estrogen? JAMA 1978;239:29-30.