ARTHUR L. HÀ SKINS, M.D. ERICA F. MOSZKOWSKI, M.D. VICTORIA P. WHITELOCK, M.D.
From the Department of Obstetrics and • Gynecology, University of Maryland School of Medicine.Presented (by invitation) at the Ninety-first Annual Meeting of the American Gynecological Society, Hot 'Springs, Virginia, May 23-25, 1968.
REVIVAL OF interest in estrioi has been occasioned by recent observations of variations in urinary. estrioi during compromised human pregnancies. In the past, research interest in estrioi had been primarily concerned with specific tissue response to the hormone. These studies produced a number of inconsistencies and contradictions con-cerning estrioi and its estrogenicity. The usual assumption that estrioi is biologically inert could not be verified, since there were ' conditions in which estrioi was more estrogenic than estradiol or estrone.
It is quite apparent that the method by which estrogenicity is evaluated is of the greatest importance. However,. it is difficult to determine which procedures indicate true estrogenicity with greatest significance. Table I indicates some of the variation in the estrogenicity of estrioi relative to the other major estrogens as reported by several investigators under ; stipulated testing' conditions. There is excellent agreement, as ex-pressed by Sealey and Sondern,1 Evans, Varney, and Koch,2 and Huffman and Groliman,3 that estrioi is less estrogenic than estradiol. Szego4 indicates to the contrary that estrioi is highly estrogenic when compared to estrone and estradiol. The difference in the observations could not be totally related to the test- conditions since, in some instances, the conditions were similar although the findings were at variance.
Table I.
The effect of the solvent on the estrogenicity of the hormone has also been at issue. As summarized in Table II, Szego, as well as Bum and Elphick,5 observed an enhancement of the estrogenicity of estrioi when the solvent was aqueous. Contrarily, Zondek and Sulman6 reported that an aqueous solvent for estrioi reduced the estrogenicity to a 10 per cent level. It is of some interest to note that in addition to technique variation, the “aqueous solvent” varied in constitution.
Table II. Estrogenicity of estriol in lipoid and “aqueous” solutions.
Discrepancies in the results of estrogenicity studies concerned with the clinical effectiveness of estrioi in humans were also noted. When, estrogenicity was determined by vaginal cytology, Table III, Brown and Bradbury7 were unable to demonstrate a positive effect with 1 mg. of estrioi daily for 10 days. Mack,8 however, with, the same dosage schedule demonstrated a positive cytologic effect similar in order to that achieved with estrone, estradiol, and stilbes- trol. Puck9 observed a positive vaginal cytologic effect with 0.1 mg. estriol daily for 5 days.
Table III. Estrogenicity of oral estriol in humans as determined by vaginal cytology.
In determining estrogenicity of estriol by withdrawal bleeding in women, Table IV, Soule10 noted this phenomenon in one patient. He compared the effectiveness of estriol as being approximately that of estradiol and stilbestrol. Puck observed that estriol given at 5 mg. a day for 6 days would not induce withdrawal bleeding. In addition to this, there was no evidence of endometrial stimulation by microscopic examination, although there was stimulation of the cervical and vaginal mucosa at this and lower dose levels.
Table IV. Estrogenicity of estriol in humans as determined by withdrawal bleeding.
One of the more interesting chapters in the estriol story is the concept advanced by Puck, in which he postulates a type of polarity attributable to estriol. This, it is suggested, results in a manifestation of estriol estrogenicity in the lower portions of the human reproductive tract (vagina and cervix) but with little or no effect upon the body of the uterus, or endometrium.
It was this concept, as well as the conflicting reports concerning the estrogenicity of estriol, that stimulated us to review the problem in our laboratories. The estrogenic effect of estriol in the immature mouse as related to uterine weight with both aqueous and lipoid solvents was to be studied. In addition,-the estrogenicity of the hormone № women was to be determined through the evaluation of the maturation index of the vaginal mucosa, ferning of the cervical mucus, estrogen withdrawal bleeding, progesterone withdrawal bleeding, and microscopic study of the endometrium after estriol administration. The estrogenic effects of estriol in women are the subject of the current presentation.
Materials and methods
Estriol, ethinyl estradiol, and stilbestrol were used throughout the study. Each was given orally. The major part of the study was concerned with estriol administration. The, other two substances were administered to small control groups of patients to provide a positive estrogenic control in addition to the negative control in the estriol group.
Women to receive estriol were selected from inpatients and outpatients at Baltimore City Hospitals and University of Maryland Hospital. The patients were selected because of evidence of ovarian failure with amenorrhea and atrophic change in the vaginal mucosa. The patients ranged in age from 47 to 94 years with the average of 66 years. Each patient received 1 mg. of estriol daily by mouth for 28 days.
Vaginal smears for cytologic study were obtained through scraping the lateral vaginal wall with a wood spatula. This procedure was followed in order to provide a standard specimen relatively free from uterine - contamination and from, exfoliated vaginal pool cells. The material obtained was smeared on a glass slide and quickly fixed in 95 per cent alcohol. Subsequently the smears were pre-pared with Shorr’s stain, the slides were then read by two investigators, and, whenever possible, 100 Cells were read and differentiated into parabasal, intermediate, and superficial cells. The cervical mucus was studied ln 30 patients, 10 of whom received estriol; 10 received 0.1 mg. of ethinyl estradiol daily for 2 - weeks and 10 patients received 0.1 Wg- of stilbestrol daily for 2 weeks. Crystal-lization of the cervical mucus was demonstrated by microscopic study.
Three patients received a synthetic pro- §estin, megestrol 5 mg. (daily for 5 days) on completion of .the estriol regimen. The purpose of this procedure was to demonstrate the potential of progestin withdrawal bleeding from the estriol-stimulated endometrium.
Two hysterectomy specimens were obtained from patients receiving estriol for 28 days prior to operation. The endometrium was examined for histologic evidence of estrogen stimulation.
Results
Estriol effects on the vaginal epithelium.
A significant change in the cell population of the vaginal epithelium was noted after 2 weeks of estriol administration (Table V). The parabasal cells decreased by approximately 50 per cent. There was a concomitant increase in the intermediate and superficial cells. The changes apparent at 2 weeks of administration were not significantly changed at the 4 weeks’ medication level.
Table V. Estriol effect on vaginal epithelium maturation index.
Comparison with stilbestrol and ethinyl estradiol is shown in Table VI. Stilbestrol at a dosage level of 0.1 mg. daily appears capable of inducing greater estrogen stimulation of the vaginal mucosa than estriol at the 1 mg. level. Ethinyl estradiol at the same dosage level as stilbestrol shows a greater estrogenic effect than stilbestrol or estriol.
Table VI. The maturation index of the vaginal epithelium after 14 days of therapy.
Cervical mucus.
In 10 patients treated with estriol at 1 mg. a day for 28 days, a positive fern test was noted in 2 patients. The remaining were negative. In similar groups of patients treated with ethinyl estradiol and stilbestrol at 0.1 mg. levels, all patients were found to have positive fern tests as indicated in Table VII.
Table VII. Ferning of cervical mucus after 14 days of therapy.
Estrogen withdrawal bleeding.
Three patients of the 60 treated with estriol showed evidence of estrogen withdrawal bleeding following therapy. This occurred within a week of the cessation of therapy and consisted of spotting for 2 to 3 days. As indicated in Table VIII, each of the patients manifesting estrogen withdrawal bleeding was noted to have stimulation of ' the vaginal mucosa at a level significantly greater than the mean of the study group.
Table VIII. The maturation index of the 3 ' patients exhibiting estrogen withdrawal bleeding.
Progestin withdrawal bleeding
Three patients were given 5 mg. of megestrol by mouth for 5 days upon the completion of the 28 day course of estriol. As indicated in Table IX, only one patient showed evidence of progestin withdrawal bleeding. This patient showed additional evidence of estro- genicity greater than that observed in the rest of the group in that the fern test was positive and the maturation. index showed an exaggerated estrogen response.
Table IX. Megestrol withdrawal bleeding.
Endometrial stimulation
Hysterectomy was accomplished on 2 patients who were in the estriol study group. Minimal estrogen stimulation of the endometrium was noted. The maturation index at the time of hysterectomy in each, respectively, was 0-84-16 and 14-75-11.
Conclusions
When given orally at a dose level of 1.0 mg. daily for 28 days, estriol produced an estrogenic effect in women. There , is stimulation of the vaginal muCosa which becomes apparent by the second week of administration and remains constant through the following 2 weeks of administration. Return to pretreatment levels of maturation of the vaginal mucosa is..noted on the smears taken 4 weeks after therapy. When compared to the vaginal cytologic effect of stilbestrol and ethinyl estradiol, it is noted that 1 mg. of estriol does not induce as great a degree of estrogenic stimulation of the vaginal mucosa as does 0.1 mg., stilbestrol. Ethinyl estradiol at this dosage level induces an intense estrogenic effect on the vaginal mucosa.
Other evidence of weak estrogenicity of estriol is seen in the occasional development of a positive fern test in the cervical mucus as compared to the 100 per cent development of positive fern tests with smaller amounts of ethinyl estradiol and stilbestrol.
Estrogen withdrawal bleeding with estriol occurred in 3 patients of the 60 study group. Withdrawal bleeding following the administration of a progestational agent occurred in one of three trials.
The studies indicate that estriol is estrogenic in the various modalities tested. The degree of estrogenicity is relatively slight and equivalent to less than 0.1 mg. , of stilbestrol.
Estriol and megestrol were supplied by Organon, Inc., and Mead Johnson & Company, respectively.
Discussion
DR. WILLARD M. ALLEN, St. Louis, Missouri. The studies reported by Dr. Haskins do show that estriol is a weak estrogen which has a specific effect on the cervix and the vaginal epithelium. A histologic effect on the endo-metrium was not detected. However, an effect must have been produced; estrogen withdrawal bleeding occurred in three trials, of 60, and progesterone withdrawal bleeding in one of three trials. The three target tissues tested—vaginal epithelium, cervical glands, and endometrium respond to estriol in the same way as they respond to other estrogens.
Estriol has been available for clinical use under the trade name of Theelol for at least 35 years. It has not been utilized very much, I suppose, because it is a “weak estrogen.” The descriptive adjective “weak” has little real significance. Let me ask a pertinent' question: Which is a better estrogen for clinical use, ethinyl estradiol or Premarin?” The first is a ighly potent estrogen, the second is a weak estr°gen, yet both in proper dosage are highly e ective and useful estrogens.
This careful study was done to see whether 0r not there was a qualitatively different effect °n the three target tissues tested. It would benice, for example, if estriol had a relatively greater effect on the vaginal epithelium and cervical glands than on the endometrium. If that were the case, oral estriol would be better for the oral treatment of atrophic vaginitis, or perhaps even the menopausal syndrome than estrogens now in use. No such difference was detected—all target tissues responded equally.
In 1944, I published a paper in the Southern Medical Journal giving the results of the bioassay of several estrogens in 8 young ovariec-tomized women with an in taetuterus. The amount of each estrogen over a 3 week period required to induce estrogen withdrawal bleeding was determined. Ethinyl estradiol given orally proved to be the most active of the estrogens tested. Also ethinyl estradiol was at least 30 times as active as estradiol. The question which now arises is, “How active is estriol?” In Haskins’ group estrogen withdrawal , bleeding occurred in 3 of 60 trials. I would predict that a dose of 3 to 5 mg. per day would produce withdrawal bleeding in 50 per cent or more of in-stances.
It seems to me, therefore, that Dr. Haskins’ study proves that estriol is an effective estrogen that might be half as active as estradiol. Manifestly, estriol will not be used clinically to treat estrogen deficiency. However, it just may be that the estriol circulating in the blood of the mother during pregnancy may have a good deal of significance.
Dr. Haskins’ paper should give impetus to the study of blood levels of E1: E2, and Es m the blood during both normal and abnormal pregnancy.
DR. R. L. VANDE WIELE, New York, New York.' I would like to make a few comments to the interesting contribution of Dr. Haskins. In analyzing the physiologic significance of his studies, it is necessary to compare the dose of estriol used by Dr. Haskins to the amonunt of estriol produced in vivo. In the nonpregnant female, 1 mg. of estriol, the dose found by Dr. Haskins to have a significant biologic effect, is an enormous amount of estriol. At the peak of estrogen secretion, the ovary produces in the neighborhood of 200 μg of estradiol per day, only a fraction of which is converted to estriol. Furthermore, it is quite likely that only insignificant amounts of free estriol, the form used by Dr. Haskins, circulate, and that most of it is conjugated before entering into the circulation.
In the pregnant female, however, the situation is very different. Available data indicate that m the third trimester somewhere between 20 and 80 mg. of estriol enters into the maternal circulation. In addition, it has been shown by Dr. Ryan, a member of our Society, that^ mosq if not all, of the estriol enters into the circulation as the free compound. It is quite likely that such high amounts of estriol have significantbiologic effects.
DR. HOWARD W. JONES, JR., Baltimore, Maryland. Dr. Allen pointed out that it would be exceedingly helpful to have an estrogen which acted on the vagina and cervix and not the endometrium. Perhaps equally important would be to have some estrogen which had a differential effect on hot flushes and did not effect endometrium.
Therefore, I rise simply to ask Dr. Haskins if he could tell us about the effect of estriol on hot flushes.
DR. EDWARD C. HUGHES, Syracuse, New York. I presented a paper entitled, “Nutritional Aspects' of the Endometrium” at the Annual Meeting of the American Association of Obstetricians and Gynecologists some years ago.I suggested that the dose level to stimulate the function of the endometrium was 0.1 mg. 0f diethylstilbestrol. I have maintained that idea for a great many years. Lately, I have proved this to myself by testing the amount of estrogen which stimulates the endometrium through organ culture of the endometrium. We find that the size of the dose is most important for the stimula-tive effects upon the metabolic activity of, the endometrium.
For instance, if you add more than 5 μg to the culture media, the uptake of glucose from the fluid is decreased, and the metabolism of glucose by the endometrium to glycogen is decreased.
If you use less than a microgram (and we have not decided how far we must go—we are down to a hundredth of a microgram right now) we do find that the uptake of glucose from the culture media does increase. The endometrium does metabolize glycogen in considerable quantities, and the enzymes- that are particularly in-volved in the synthesis, phosphorylase and synthetase, are likewise stimulated.
If you use less than a microgram (and we have not decided how far we must go—we are down to a hundredth of a microgram right now) we do find that the uptake of glucose from the culture media does increase. The endometrium does metabolize glycogen in considerable quantities, and the enzymes- that are particularly in-volved in the synthesis, phosphorylase and synthetase, are likewise stimulated.I think your suggestion, that a small dose of estrogen has a stimulative effect upon the generative tract, is very important. If you wish to depress the action, particularly m the endometrium, a larger dose is appropriate.
DR. HASKINS (Closing). In regard to question concerning 0.1 mg. stilbestrol: we have for a number of years in the treatment of patients with ovarian agenesis administered 0.1 mg. stilbestrol daily without interruption and with 5 days a month of added synthetic progestin This will produce endometrial stimulation an evoke regular menstrual bleeding.
The symptom of hot flushes was not studie • Indeed, the majority of our patients di no offer this complaint. This finding was undoubtedly relative to the advanced age of thepatient population.
The question of the glucuromde or og the conjugation making the estriol inactive, I think is certainly an excellent point, but one w ic am unable to discuss competently.
References
Sealey, J. L., and Sondern, C. W.: Endocrinology 29: 356, 1941.
Evans, J. S., Varney, R. F., and Koch, F. C.: Endocrinology 28: 747, 1941.
Huffman, M. N., and Grollman, A.: Endocrinology 41: 12, 1947.
Szego, C. M.: Fed. Proc. 9: 124, 1950.
Burn, J. H., and Elphick, G. K.: Quart. J. Pharmacol 5: 192, 1932.
Zondek, B., and Sulman, F.: Endocrinology 49: 417, 1951.
Brown, W. E., and Bradbury, J. T.: J. Clin. Endocrinol. 9. 725, 1949.
Mack, H. C.: AM. J. OBST. & GYNEC. 45: 402, 1943.
Puck, A.: München, med. Wchnschr. 99: 1 1957.
Soule, S. D.: AM. J. OBST. & GYNEC. 44: 648, 1942