The relative potencies of commercially available estrogens are of great importance when prescribing estrogen, and the clinician should be familiar with the following potencies:

The 17α-ethinyl group of ethinyl estradiol (by resisting metabolism) enhances hepatic effects, because no matter by which route it is administered, liver function is affected.¹³ The same is true for conjugated equine estrogens. Contrary to the case with estradiol, the liver appears to preferentially extract ethinyl estradiol and conjugated equine estrogens no matter what the route of administration. Thus, the route of administration appears to influence the metabolic responses only in the case of specific estrogens, most notably estradiol.

A major factor in the potency differences among the various estrogens (estradiol, estrone, estriol) is the length of time that the estrogen binds to its receptor. The higher rate of dissociation with the weak estrogen (estriol) can be compensated for by continuous application to allow prolonged binding and activity. Estriol has only 20-30% affinity for the estrogen receptor compared with estradiol; therefore, it is rapidly cleared from a cell. However, if the effective concentration is kept equivalent to that of estradiol, it can produce a similar biologic response.¹⁶ At least two studies have been unable to demonstrate prevention of bone loss with the administration of 2 mg estriol daily.^17,18 In pregnancy, where the concentration of estriol is very great, it can be an important hormone not just a metabolite. Thus, higher estriol levels are not necessarily protective against potent estrogenic effects. Because estriol protects the rat against breast tumors induced by various chemical carcinogens,¹⁹ it has been hypothesized that a higher estriol level protects against the more potent effects of estrone and estradiol. But, antagonism of estradiol occurs only within a vary narrow range of the ratio of estradiol to estriol, a range that is rarely encountered either physiologically or pharmacologically.²⁰ Below this range, estradiol is unimpeded, above this range estriol itself exerts estrogenic activity. The commercial preparation that contains estriol, estradiol, and estrone contains sufficient amounts of estrone and estradiol to produce standard clinical effects.

Esterified estrogens are synthetically prepared from plant precursors and are composed mostly of sodium estrone sulfate with a 6-15% component of sodium equilin sulfate. Estradiol valerate is rapidly hydrolyzed to estradiol; therefore, the pharmacology and effects are comparable at similar dosages.²¹

The patches first used for transdermal estrogen administration contained an alcohol reservoir; the estrogen was released through a semipermeable membrane attached to the skin with an adhesive. In the current generation of patches, the hormones are dissolved and distributed throughout the adhesive matrix. In a study of women who had previously discontinued patches because of skin irritation (contact dermatitis), skin reactions were less common with the newer matrix patches.²² In addition, the matrix patches are better tolerated in tropical environments.²³ The patches are designated according to the amount of estradiol delivered per day: from 14 to 100 μg.

The concentration of estrogen in the hepatic portal system after oral administration is 4-5 times higher than that in the periphery.²⁴ Because of first-pass metabolism in the liver, oral estradiol results in a circulating estrone to estradiol ratio of approximately 3; with transdermal administration the ratio is 1. The first-pass effect may be important for lipoprotein effects. For example, short-term studies (6 weeks) could document increased catabolism of low-density lipoprotein (LDL-cholesterol) and increased production of apoprotein A-I with oral estrogen, but no effect with transdermal estrogen.^25,26 A 2-year study in Los Angeles with a transdermal dose (100 μg) detected no significant change in HDL-cholesterol levels.²⁷ However, English data indicate that the transdermal administration of 50 μg estradiol twice a week is as effective as 0.625 mg oral conjugated estrogens, when combined with a progestin in sequential regimens, on bone density and lipids over a duration of 3 years.²⁸ Standard doses of estrogen administered transdermally (50 μg) protect against fractures as well as standard oral doses do.²⁹ As with oral estrogen, lower transdermal doses can produce effects on bone density and menopausal symptoms,³⁰ but a substantial number of women require higher doses.

The critical question is whether the first-pass effect of oral estrogen is clinically important. The different effects of oral and transdermal administration on metabolic parameters have been repeatedly compared over the years, but epidemiologic studies of clinical end points are not abundant, handicapped by the relatively small numbers of women using transdermal estrogen in most countries.

Clotting Factors. First-pass hepatic metabolism affects the synthesis of clotting proteins, markers of coagulation and fibrinolysis that can influence the risk of thrombosis and coronary heart disease events. Oral estrogen increases factor VII and prothrombin 1 and 2 fragment, whereas transdermal estrogen decreases factor VII.^31,32,33 and ³⁴ Oral estrogen also increases circulating levels of matrix metalloproteinases, MMP-2 and MMP-9, enzymes that are associated with a tendency for clotting.³⁵ However, what is important is whether the different effects of oral and transdermal delivery on clotting factors translate into clinical differences and cardiovascular risk.

Activated Protein C (APC) Resistance and Risk of VTE. Resistance to APC is an important marker for venous thrombosis in individuals with inherited thrombogenic mutations and even in the absence of these mutations. Oral estrogen increases APC resistance, whereas transdermal estrogen has no significant effect on this marker.^36,37 Based on this difference, one would predict that transdermal delivery of estrogen would be less likely than oral delivery of estrogen to be associated with venous thromboembolism (VTE).

A French case-control study (epidemiologic studies of the link between the transdermal route of administration and a relatively rare event are possible in France because of the popularity of the transdermal method) reported no increased risk of VTE in users of transdermal estrogen, as compared with a 4-fold increase in oral estrogen users.^38,39 and ⁴⁰ Estrogen users who carried a factor V Leiden mutation or a prothrombin mutation had a 25-fold higher risk of VTE than did women who did not use estrogen and did not have either mutation. The women with a prothrombotic mutation who used transdermal estrogen had a VTE risk that was similar to that of women with a prothrombotic mutation who did not use estrogen. The French E3N
prospective cohort study also reported an increased risk of venous thromboembolism with current users of oral therapy, a hazard ratio of 1.7 (CI=1.1-2.8), a ratio that is similar to the usual 2-fold increase repeatedly documented in the literature, and no increase with transdermal estrogen.⁴¹ Venous thrombosis is discussed in more detail later in this chapter.

Lipids and Hepatic Enzymes. Both oral and transdermal estrogen reduce total cholesterol, low-density lipoprotein cholesterol, and lipoprotein(a). Compared with transdermal estrogen, oral estrogen produces significantly greater elevations in high-density lipoprotein cholesterol and increases triglycerides, whereas transdermal estrogen decreases triglyceride levels.^31,33,42,43 and ⁴⁴ Indeed, triglyceride levels markedly elevated in response to oral therapy return to normal when treatment is changed to transdermal administration.⁴⁵

Inflammatory Markers. Women on oral estrogen have increased levels of C-reactive protein (CRP), whereas those taking transdermal estrogen do not.^{31,33,42,46,47,48} and ⁴⁹ In fact oral hormone therapy while increasing CRP, as discussed in Chapter 17, reduces the circulating levels of other inflammatory markers (E-selectin, intercellular adhesion molecule-1, vascular cell adhesion molecule-1, monocyte chemoattractant protein-1, and tumor necrosis factor-α) with inconsistent effects on interleukin-6.^46,47 Transdermal estrogen does not affect levels of these inflammatory markers. It is not certain that the decrease in CRP levels with statins and the increase with oral estrogen are instrumental in clinical outcomes or reflect other effects. Thus raising or lowering CRP levels will not necessarily increase and decrease the risk of clinical disease.

A longitudinal study of 346 postmenopausal women taking oral hormone therapy reported that elevated CRP was a strong predictor of future cardiac events, but only in those with increased IL-6 levels.⁴⁸ An increase in CRP alone was not associated with an excess of events. The difference in CRP levels between users of oral versus transdermal therapy, especially in younger postmenopausal women, is of little clinical significance. In fact, in the Estrogen Replacement on Progression of Coronary Atherosclerosis trial, estrogen-induced increases in CRP had no effect on disease progression, as measured by serial angiograms.⁴⁹ A study from the Women’s Health Initiative confirmed the correlation between baseline levels of CRP and an elevated risk of coronary heart disease, but the increase in CRP induced by oral hormone therapy did not further increase the risk!⁵⁰

Myocardial Infarction Risk. Both oral and transdermal administration of hormone therapy are associated with a decrease in myocardial infarction risk in observational studies.⁵¹

Metabolic Syndrome. In a 3-month randomized trial involving 50 obese women with metabolic syndrome, oral estradiol therapy worsened markers of the metabolic syndrome, including insulin resistance, suggesting a worsening of cardiovascular risk, whereas transdermal estradiol had minimal effects.⁵²

Effects in Smokers. Limited evidence suggests that postmenopausal women who smoke may have a better cardiovascular response to transdermal estrogen than to oral estrogen, including greater reductions in total peripheral resistance, vascular sympathetic tone, and norepinephrine levels, and increased vascular responsivity.⁵³ Smokers receiving transdermal estradiol have decreased plasma viscosity and thromboxane B₂ levels.⁵⁴ These results raise the possibility, although the data are limited, that smokers may represent a group of women for whom transdermal estrogen would be an advantage.

Carbohydrate Metabolism. There is little difference between the oral and transdermal methods of delivery on carbohydrate metabolism. Both methods have a beneficial impact on central abdominal fat content, glucose levels and insulin resistance, associated with a reduced risk of developing adult-onset diabetes mellitus.^55,56,57,58 and ⁵⁹

Breast Cancer Risk. Oral conjugated equine estrogens/sequential medroxyprogesterone acetate decreased median levels of insulin-like growth factor-1 (IGF-1) by 26% and
increased median levels of sex hormone-binding globulin (SHBG) by 96% relative to baseline, whereas no change occurred with transdermal estradiol.⁶⁰ High IGF-1 and low SHBG levels are associated with increased breast cancer risk; however, it is difficult to make clinical conclusions based on these secondary markers. A German case-control study of 3,593 cases found no significantly increased risk of breast cancer with oral or transdermal hormone therapy.⁶¹ Thus far, the epidemiologic data comparing oral and transdermal treatment are not sufficient to allow firm conclusions regarding breast cancer risk.

Colorectal Cancer Risk. In a case-control study, both oral and transdermal hormone therapy reduced the risk for developing colorectal cancer.⁶² When transdermal therapy involved estrogen alone, the benefit was even greater.

Estradiol Levels in Users of Oral Versus Transdermal Estrogen. Studies comparing circulating estradiol levels in women receiving oral or transdermal estrogen reveal therapeutic estradiol levels predictive of a good bone response, but they also contain large standard deviations, indicating substantial variation among individuals.⁶³ Individual women metabolize estrogen differently, depending on the route of administration, liver function, skin absorption, body composition, body size, potential medication interactions, and the presence of binding proteins; all of which contribute to individual variations in serum estradiol levels.⁶⁴ In the future, measurement of serum estradiol levels may play a role in assessment of adequate treatment. This measurement will be especially useful for users of transdermal estrogen therapy, which produces more consistent estradiol levels than does orally administered therapy.

The only way to accurately compare clinical differences between oral and transdermal estrogen delivery is to establish that the two methods produce similar blood levels and that clinical differences reflect the first-pass effect through the liver. This is difficult to accomplish because the oral first-pass effect raises sex hormone-binding globulin (SHBG) levels such that total serum estradiol levels are greatly affected. A study of 18 women showed that oral estrogen increased SHBG by 67% to 171%, whereas transdermal estrogen did not alter SHGB levels.⁶⁵ Estrogen-induced changes in SHBG may be clinically significant because estrogen unbound to SHBG determines the estrogen effects of a given regimen. The only study that measured free estradiol levels, compensating for increases in SHBG, indicated at 12 weeks, that serum free estradiol levels in the oral group were similar to those in the transdermal group.³² However, because these results were derived from only 18 women, the effect of oral and transdermal doses on free estradiol levels has not been reliably established. A potential advantage of transdermal treatment because it has no effect on SHGB levels is the absence of a reduction in free, unbound testosterone levels as is observed with oral therapy.⁶³ Thus, the transdermal method may be indicated in women with impaired sexual function.

Oral Versus Transdermal Administration. It is difficult to draw conclusions about clinical differences between oral and transdermal hormone delivery based on secondary markers. Epidemiologic studies on clinical events are needed. However, this is a challenge because of the relatively small number of women receiving transdermal estrogen. In addition, the studies must adjust for individual variability of dosing to ensure that circulating estrogen levels in the patients being studied are similar.

Some patients do not gain full relief from the symptoms of vaginal atrophy with oral or transdermal administration of estrogen. Local vaginal administration makes sense for these patients. Vaginal treatment is especially helpful when a rapid response is desired. In addition, there are many women who desire the genitourinary effects of estrogen but either must or wish to avoid systemic therapy. Overall, there is no evidence that one method or preparation is superior to the others in achieving clinical response. Measurement of vaginal pH from the lateral vaginal wall is a simple and inexpensive way to assess adequate treatment of the vagina. It has been impressive in our experience and others how an acidic pH (<4.5) obtained from the lateral, outer third of the vagina correlates well with good estrogen effects.^66,67 and ⁶⁸

Many clinicians believe that estrogen administered intravaginally is not absorbed, and systemic effects can be avoided. This is not the case. Estrogen in creams is absorbed very readily from a vagina with immature, atrophic mucosa.⁶⁹ Indeed, the initial absorption is rapid, and relatively high circulating levels of estrogen are easily reached. As the vaginal mucosa matures, absorption decreases.⁷⁰ This decline takes approximately 3-4 months, after which lesser but still significant absorption takes place.⁷¹ Effective treatment of vaginal atrophy with minimal absorption can be achieved with the administration of 0.3 mg conjugated estrogens, 2-3 times per week.^72,73 We believe that treatment with a vaginal cream longer than 6-12 months requires endometrial surveillance.

The amount of estradiol delivered in low-dose tablet form or a ring is not sufficient to treat menopausal symptoms, but effectively improves local urogenital atrophy and reduces recurrent urinary tract infections. This has been accepted as a method to relieve atrophic vaginal symptoms in women with contraindications to estrogen treatment; however, systemic effects do occur.

Estring is a 55-mm diameter silicone ring that contains 2 mg estradiol, with a release rate of 7.5 μg/day for 90 days.⁷⁴ European studies have demonstrated that vaginal maturation can be achieved with this ring that can be left in place for 3 months, with a low-level of systemic absorption.^75,76 The subjective symptoms associated with vaginal atrophy are rapidly relieved. No change in endometrial thickness was observed after 1 year of treatment.⁷⁷

Vagifem is a tablet that contains 25 μg estradiol, and the initial dose of 1 tablet daily produces relief from atrophic symptoms within 2 weeks.⁷⁸ After the first 2 weeks, the maintenance dose is twice weekly, and endometrial thickness has been reported to not change from baseline; however, the study was only 6 months in duration.⁷⁹ One 2-month study found no evidence of endometrial stimulation; another reported 1 case of vaginal bleeding with endometrial proliferation.^80,81 A smaller dose tablet, 10 μg, also improves vaginal atrophy, but it is not as effective as the larger dose.⁸²

The systemic absorption of estrogen from the low-dose estradiol ring or tablet is very low, especially after the vagina achieves estrogen-induced maturation (about 3 months). Is this low level of absorption free of the risk of endometrial hyperplasia? The problem is that all studies
have been too short (all 1 year or less, except one 2-year study) to determine long-term endometrial safety. Although systemic absorption occurs, the circulating estradiol levels with these low-dose methods remain in the normal postmenopausal range.^{74,79,83,84,85} and ⁸⁶ But, because the small increase in circulating estradiol levels causes distant target tissue responses (e.g., an increase in bone density or an improvement in the lipid profile^87,88), clinicians cannot assure patients that these methods are totally free of systemic activity. Although the change in blood levels is very slight, and for that reason not effective for the relief of vasomotor symptoms, we believe long-term treatment requires ultrasonographic monitoring of endometrial thickness with biopsy when indicated. This ultrasonographic approach is more preferable than complicating the treatment regimen with the addition of a progestational agent. We further suggest that each patient titrate her dose and schedule of treatment to balance an effective response with minimal dosing. For women who are breast cancer survivors and are considering this treatment, clinicians and patients must accept a small but real unknown risk.

A vaginal ring (FemRing, MenoRing) that releases estradiol acetate provides 50 or 100 μg estradiol per day over a 3-month time span.^89,90 Estradiol acetate is a prohormone, which is rapidly hydrolyzed to estradiol that is reflected in blood estradiol levels similar to those achieved with oral and transdermal methods. The systemic levels achieved effectively suppress hot flushing, and a beneficial impact on bone is to be expected. Endometrial protection requires the addition of a progestin in the presence of a uterus.

Estradiol pellets are available in doses of 25, 50, and 75 mg for subcutaneous administration twice yearly. The 25-mg pellet provides blood levels in the range of 40-60 pg/mL, levels that are comparable with those obtained with standard oral doses.^91,92 However, the effect is cumulative, and after several years the blood levels are 2-3 times higher. Significant blood levels of estradiol will persist for up to 2 years after the last insertion. We believe that the estradiol pellets confer no advantages over the usual treatment regimens. We recommend that women receiving pellets be monitored with blood estradiol levels, and levels greater than 200 pg/mL (and preferably, 100 pg/mL) should be avoided by a greater interval between insertions.

Transdermal estradiol can also be administered by a gel, emulsion, or spray. The gel, available in various trade names (Divigel, Elestrin, Estrogel, Estreva Gel), is applied by a metered pump or from a foil packet once daily on an arm, anywhere from the wrist to the shoulder, or the thigh, without rubbing or massaging and alternating sides.^93,94 The emulsion, Estrasorb, is packaged in foil pouches; usually two packets are applied daily, one to each thigh, and rubbed in thoroughly. Evamist is the transdermal spray, and the usual dose is one spray daily to the forearm (if more than one dose is required daily, each spray is on a separate site).⁹⁵ Simultaneous use of sunscreen on the site of administration should be avoided. If dosage is being monitored by blood estradiol levels, blood should be drawn from a site where transdermal estradiol has not been applied for several days. Although comparison studies have not been performed, it is reasonable to expect similar pharmacokinetics for all transdermal methods. As with pellets, we recommend that blood estradiol levels be monitored and maintained at a level below 100-200 pg/mL.

Monitoring the estradiol blood level in postmenopausal women receiving hormone therapy is not as straightforward as it would seem. There are two primary difficulties. First, the clinical assays available differ considerably in their technique and quality (laboratory and antibody variations). Second, the various commercial products represent a diverse collection of estrogenic compounds, ranging from estradiol to unique equine estrogens. Although the body interconverts various estrogens into estrone and estradiol, is this process relatively consistent within and between individuals? A highly specific assay for estradiol will detect very low levels of estradiol in women receiving 0.625 mg conjugated equine estrogens; nevertheless, most clinical assays will report a level of 40-100 pg/mL in these women.

We find measurement of blood estradiol levels very useful in selected patients, such as the patient who requests ever-increasing doses of estrogen for the treatment of symptoms, which in the presence of very high blood levels of estradiol can be confidently diagnosed as psychosomatic. We further advocate titering of estrogen dosage with blood estradiol levels in women who fail to demonstrate a positive bone response on treatment, as discussed in Chapter 17. What each clinician must do is learn what blood level of estradiol as performed by the local laboratory is associated with the standard doses of hormone therapy (0.625 conjugated estrogens, 1 mg estradiol, 50 μg transdermal estradiol) and consistently use the same laboratory. In our laboratory this range is 40-100 pg/mL estradiol when the estrogen is taken the evening before the office visit (with transdermal administration blood sampling should be obtained the day before new patch placement); the range reflects individual variation including the variability from peak to nadir values. Remember that because FSH is regulated by a factor other than estrogen (i.e., inhibin), FSH levels cannot be used to monitor estrogen dosage. Postmenopausal hormone therapy will produce only a 10-20% decrease in FSH and LH, and there is great individual variability in the responses.⁹⁶

Products containing ethinyl estradiol will not affect the measurement of circulating estradiol levels. Ethinyl estradiol circulates without being changed, and the antibodies in the immunoassays for estradiol will not recognize it. It is for this reason that women on oral contraceptives have very low measurements of estradiol. This problem for the postmenopausal use of ethinyl estradiol is not a major handicap because ethinyl estradiol is slowly metabolized, and blood levels are relatively stable with less variation from individual to individual compared with the other estrogen formulations.

Postmenopausal hormone therapy initially consisted only of sequential regimens that were logical reflections of the cyclic estrogen and progesterone patterns in a premenopausal menstrual cycle. Clinical trials established the doses and durations for progestin administration that would effectively protect the endometrium against unchecked proliferation.⁹⁷ Progestin withdrawal bleeding occurs in 80-90% of women on a sequential regimen,^98,99 and ¹⁰⁰ and for this reason the continuous combined method of treatment evolved to improve patient continuance that was adversely affected by bleeding and other symptoms triggered
by the cyclic hormonal changes. The addition of a daily dose of a progestin to the daily administration of estrogen allowed the progestin dose to be smaller, provided effective protection against endometrial hyperplasia, and resulted in amenorrhea within 1 year of treatment in 80-90% of patients.^99,101,102 and ¹⁰³

In the sequential regimen, estrogen is administered daily and progestins for 2 weeks of every month, using the comparable doses of the following progestins^{100,101,104,105}:

5 mg medroxyprogesterone acetate, or 0.7 mg norethindrone, or 1.0 mg norethindrone acetate, or 200 mg micronized progesterone.

In the daily continuous, combined regimen, progestins are combined with estrogen in the following comparable doses^102,103,106:

1.5 or 2.5 mg medroxyprogesterone acetate, or

0.35 mg norethindrone, or

0.5 or 1.0 mg norethindrone acetate (0.1 mg dose is available), or

100 mg micronized progesterone or

2 mg drospirenone or

2 mg dienogest.

There has been a progressive decrease in dose used for postmenopausal hormone therapy. For many years, the standard dose of estrogen was 0.625 mg conjugated estrogens, 1-2 mg micronized estradiol, 1-2 mg estradiol valerate, or equivalent doses of other estrogens such as 5 μg ethinyl estradiol. Lower doses have been proven on the average to be as effective as these “standard” doses, providing clinicians and patients with more options. Conjugated estrogens in a dose of 0.3 or 0.45 mg effectively produce a gain in bone density when combined with 1.5 mg medroxyprogesterone acetate, and a dose of 0.5 mg micronized estradiol produces comparable effects.^107,108,109 and ¹¹⁰ The 0.45/1.5 mg and 0.3/1.5 mg conjugated estrogens/medroxyprogesterone acetate combinations improve vaginal atrophy, reduce hot flushing, and improve measures of sexual function in a pattern that is quantitatively and qualitatively similar to the 0.625/2.5 mg combination with less mastalgia.^111,112 These lower-dose combinations are associated with less breakthrough bleeding and a higher rate of cumulative amenorrhea compared with older standard doses and retain the favorable changes in the lipid profile.^113,114 At these lower doses of conjugated estrogens, the combination with progestin produces an additive effect; therefore, when these lower doses of estrogen are used without progestin, the effect on hot flushing will not be as great. In a dose-response study, the most efficacious dose of oral micronized estradiol was 1 mg/day.¹¹⁵ The lower-dose combination of ethinyl estradiol and norethindrone acetate (2.5 μg/0.5 mg) is nearly as effective in treating hot flushes as the higher dose combination (5.0 μg/1.0 mg).¹¹⁶

Two metabolites of progesterone, allopregnanolone and pregnanolone, are believed to be responsible for progesterone’s unique sedative effect. Treatment regimens with micronized progesterone should be taken at bedtime, and these estrogen-progesterone combinations are a good choice for women with sleep difficulties. A study in a sleep laboratory has demonstrated a significant improvement in sleep quantity and quality in women using a
sequential regimen of estrogen and micronized progesterone in contrast to no effect in the group using medroxyprogesterone acetate.¹¹⁷

Many women do not tolerate treatment with progestational hormones. Typical side effects include breast tenderness, bloating, and depression. These reactions are significant detrimental factors with continuance. However, appropriately designed, placebo-controlled studies fail to document adverse physical or psychological effects with short-term treatment utilizing medroxyprogesterone acetate, except for breast discomfort.^118,119,120 and ¹²¹ This suggests that progestin side effects other than mastalgia are related to duration of treatment or that only studies with large numbers of subjects will detect the small percentage of women who have problems (and both explanations are probably true).

Breast discomfort associated with postmenopausal hormone therapy can be attributed largely to progestins. In the PEPI randomized trial, an increase in mastalgia was observed only in 28.7% of the women receiving estrogen-progestin combinations, containing either medroxyprogesterone acetate or progesterone.¹²⁰ Comparison studies have not been performed to address whether this symptom is minimized by particular progestins. It has been our experience, that changing to a regimen containing norethindrone or norethindrone acetate has been beneficial (but this may reflect either a placebo response or diminishing severity with time).

Can the progestational agent be administered less frequently? We are secure in our position, supported by clinical data, that a daily combination program effectively prevents endometrial hyperplasia. A sequential regimen that incorporates progestin exposure for less than 14 days has over time an increased risk of endometrial hyperplasia.^122,123 In a Finnish study, sequential regimens in standard schedules used for at least 5 years were associated with an increased risk of endometrial cancer.¹²⁴ Thus, sequential regimens with less than 14 days of progestin monthly or even long-term use of recommended schedules do not match the protection offered by the daily, continuous method of estrogen-progestin treatment.

Experience with extended cycle regimens is very limited. The administration of medroxyprogesterone acetate every 3 months was associated in 1 study with longer, heavier menses and unscheduled bleeding and a 1.5% incidence of hyperplasia at 1 year, whereas in another study, overall bleeding was less, but the incidence of hyperplasia was approximately 4%.^125,126 In a Dutch study that was only 12 weeks in length, simple endometrial hyperplasia was encountered at the end of the unopposed estrogen phase.¹²⁷ In yet another study, there was no endometrial hyperplasia encountered by 143 women who completed 2 years of treatment; however, the progestin administered every 3 months was of high dosage, 20 mg medroxyprogesterone acetate daily for 14 days.¹²⁸ In Finland, the addition of progestin at 3-month intervals was associated with a striking increase in the risk of endometrial cancer when this regimen was used for many years.¹²⁴ Most impressively, the Scandinavian Long Cycle Study, a clinical trial scheduled to last 5 years, was canceled after 3 years because of a 12.5% incidence of endometrial pathology and 1 case of endometrial cancer.¹²⁹ Therefore, if a patient chooses an extended cycle regimen, endometrial monitoring is required. In our view, an annual endometrial biopsy is strongly recommended in estrogen users exposed only intermittently to progestin treatment. Any program that differs from the standard regimen is untested by clinical studies of sufficient length and patient numbers and, therefore, requires periodic surveillance of the endometrium. Even the long-term use of standard sequential regimens is subject to a small increase in the risk for endometrial cancer, and endometrial surveillance should be considered in women using this method.

Some patients are very sensitive to medroxyprogesterone acetate. In our experience, these patients are often relieved of their symptoms by switching to norethindrone. In a sequential regimen, the dose of norethindrone is 0.7 mg (available in the progestin-only, minipill oral contraceptive; each pill contains 0.35 mg norethindrone). In the continuous, combined regimen, the dose of norethindrone is 0.35 mg daily. Commercial combination products are available containing estradiol and norethindrone acetate.

Progesterone can be administered in a vaginal gel that allows the delivery of very low doses that can effectively protect the endometrium with low systemic levels because of a first-pass effect on the uterus.¹³⁰ The administration of 90 mg every 2 days produces secretory changes in the endometrium.¹³¹ An application of the 4% commercial preparation twice weekly protects the endometrium and is associated with amenorrhea in most patients. In a sequential regimen, the 4% preparation should be applied daily for at least 14 days each month. No long-term studies are available that document endometrial safety and metabolic effects.

The transdermal estrogen-progestin combinations incorporate norethindrone acetate in a daily dose of 0.140 or 0.250 mg; or levonorgestrel in daily doses of 0.007, 0.015, 0.030, and 0.040 mg/day; and in a sequential regimen, norethindrone acetate, 0.250 mg, or levonorgestrel, 0.010 mg.^132,133 and ¹³⁴

The contraceptive levonorgestrel-releasing intrauterine system (IUS) has been reconfigured in a smaller model (not yet available) that releases 10 μg of levonorgestrel per 24 hours; however, the larger, contraceptive levonorgestrel IUS (Mirena) can also be used in postmenopausal women.^{135,136,137,138} and ¹³⁹ The intrauterine presence of the progestin effectively protects the endometrium against hyperplasia and cancer.¹⁴⁰ The local site of action provides endometrial protection and escapes systemic progestin side effects; for example, estrogen’s favorable lipid effects are not attenuated.¹⁴¹ As with the oral continuous, combined regimens, there is irregular breakthrough bleeding in the first 6 months, and after 1 year, approximately 60-70% of the women are amenorrheic. The levonorgestrel system has the advantage of a 10-year duration of use. The frameless IUD has also been designed for postmenopausal use (FibroPlant-LNG), delivering 14 μg of levonorgestrel per 24 hours.^{142, 143} These methods provide treatment options that minimize, if not totally eliminate, the systemic effects of progestins. See Chapter 25 for a complete discussion of advantages and problems.

There are some special conditions that warrant the use of a combined estrogen-progestin regimen in hysterectomized women.

Raloxifene exerts no proliferative effect on the endometrium but produces favorable responses in bone and lipids.^180,181,182 and ¹⁸³ The MORE (Multiple Outcomes of Raloxifene Evaluation) study of raloxifene administration to osteoporotic women reported results from 8 years of follow-up.^184,185 Women with low T-scores had approximately a 50% reduction in vertebral fractures with raloxifene treatment, and with previous vertebral fractures, approximately 35%. However, there has been no evidence of a reduction in hip or wrist fractures. The major side effect was about a 3-fold increase in venous thromboembolism. Raloxifene (and tamoxifen) share with estrogen an increased risk for venous thromboembolism.¹⁸⁶ The size of the risk is comparable for all three drugs, and nearly all the cases occur in the first 1 or 2 years of exposure. A small number of women experience hot flushing with raloxifene. Raloxifene treatment in the MORE trial had neither a positive nor a negative effect on cognition.¹⁸⁷

Women who received raloxifene in the MORE trial had about an 80% reduction in the incidence of estrogen receptor-positive breast cancers. The CORE study, the Continuing Outcomes Relevant to Evista trial, was designed to measure the impact of four additional years of raloxifene (60 mg/day), beginning during the fourth year of the MORE trial.¹⁸⁸ Of the 7,705 participants initially randomized in the MORE trial, 3,510 women elected to continue raloxifene treatment (2,336 completed the CORE trial) and 1,703 continued on placebo (1,106 completed the trial). During the 4-year CORE study, raloxifene treatment was associated with a 66% (HR=0.34; CI=0.18-0.66) reduction of estrogen receptor-positive invasive breast cancer in the treated group. There was no difference in estrogen receptor-negative tumors. Over the entire 8-year period, the reduction in estrogen receptor-positive cancers reached 76%. In the 8-year period, there was no difference in the number of deaths in the two groups.

The Study of Tamoxifen and Raloxifene (STAR) trial enrolled 19,747 women at increased risk of breast cancer who were randomized to treatment with either raloxifene, 60 mg daily, or tamoxifen, 20 mg daily, in more than 500 centers in the U.S., Canada, and Puerto Rico.¹⁸⁹ After an average treatment period of almost 4 years, the numbers of invasive breast cancers were identical in the two groups of women. It was estimated that these results were equivalent to about a 50% reduction (based on the previous results in the tamoxifen prevention trial),^190,191 but without a placebo arm, an accurate assessment was impossible. Thus, raloxifene appears to achieve the same reduction as tamoxifen in invasive breast cancers with a lesser increase in venous thrombosis, and perhaps no increase in cataracts and uterine cancer. Fractures, as well as strokes and heart attacks, were equally prevalent in the two groups. “Quality of life” was said to be the same for both drugs.

Tamoxifen has been demonstrated to reduce the incidence of both lobular carcinoma-in-situ and ductal carcinoma-in-situ.^190,191 In the 7-year follow-up report of the tamoxifen for prevention study, the risk for breast cancer was 0.57 (CI=0.46-0.79), a 43% reduction, not the 50% cited in the results above, and the risk for in-situ disease was 0.63 (CI=0.45-0.89), a 37% reduction.¹⁹⁰ Not only did raloxifene not yield a reduction in in-situ cancers, the number with raloxifene in the STAR trial was greater. If raloxifene is truly preventing breast cancer, this should produce a reduction in in-situ disease. Perhaps with longer follow-up, a difference between the two treatment groups will no longer be apparent.

In a 2-year randomized trial in monkeys, raloxifene exerted no protection against coronary artery atherosclerosis despite changes in circulating lipids similar to those achieved in women.¹⁹² The Raloxifene Use for the Heart (RUTH) study included more than 10,000 women from 26 countries, either at high risk for myocardial infarction or with known coronary heart disease.^193,194 The participants were randomized to placebo or raloxifene, 60 mg daily, and followed for up to 7 years. There was no effect of raloxifene treatment on coronary heart disease events; however, there was a small increase in stroke mortality.

The results of the RUTH trial are not surprising. The known favorable impact of raloxifene on the cholesterol-lipid profile was not robust enough to prevent coronary events.

Because the Women’s Health Initiative and the Nurses’ Health Study reported a reduction in coronary events associated with estrogen therapy administered to young postmenopausal women, the RUTH trial performed a post hoc analysis of the impact of raloxifene according to age of the women at entry to the study as well as subgroups such as the use of medications, including statins.¹⁹⁴ Overall, raloxifene did not increase or decrease coronary events in either of the treated groups. The only category demonstrating a significant difference, a 40% reduction in coronary events, consisted of women less than 60 years of age. Despite the statistically significant reduction in coronary events in women under age 60, there was no relationship in any subgroup according to years since menopause, even in the group less than 10 years postmenopausal. The women who were less than 60 years of age were an average of 9.9 years since menopause, compared with 19.4 years for the overall study population. The finding of a beneficial effect of Raloxifene in the youngest postmenopausal women in the RUTH trial does not jibe with the failure to find a relationship with years since menopause. Out of the 10,101 women, there were only 360 to 460 women (14-18%) in each of the patient groups who were under age 60, and only 134 women younger than age 60 experienced a coronary event. Only one subgroup demonstrated a statistically significant different conclusion than the overall finding, out of 51 analyzed subgroups. A decision to use raloxifene should not be influenced by its effects on the cardiovascular system. This is a bone and breast decision.

Arzoxifene is an estrogen agonist-antagonist similar to raloxifene, originally studied for the treatment of breast cancer. Preclinical studies indicated that arzoxifene is an estrogen agonist in bone and on lipids, but an estrogen antagonist in endometrial and breast tissue. Arzoxifene, therefore, had the potential to be as effective as tamoxifen but be free of the risk of endometrial stimulation, and perhaps, venous thrombosis.

A Phase III clinical trial comparing arzoxifene and tamoxifen in the treatment for advanced local breast cancer or metastatic tumors was disappointing.¹⁹⁵ The trial was terminated when it became apparent that the results with arzoxifene were inferior to tamoxifen. Two other members of this drug family, droloxifene and idoxifene, have also failed to yield superior results to tamoxifen for the treatment of breast cancer. For this reason, attention was turned to another use for these agents.

A 2-year, randomized trial compared the bone density responses in 331 postmenopausal women treated with either arzoxifene, 20 mg/day, or placebo.¹⁹⁶ Bone density was slightly increased in the spine and the hip in the treated group compared with placebo. There was no evidence of endometrial stimulation in the treated group either on biopsied specimens or by measurement of endometrial thickness by transvaginal ultrasonography. Three patients in the placebo group and none in the treated group developed breast cancer. However, the initial results from a projected 5-year, phase III trial of 9,354 postmenopausal women indicated an increase in venous thromboembolic events, endometrial changes, and hot flushes, and no decrease in nonvertebral fractures in the treatment arm.¹⁹⁷ The overall impact was deemed insufficient to achieve a competitive edge, and development for osteoporosis prevention and treatment was terminated.

Ospemifene (Ophena), another estrogen agonist-antagonist, is equally effective as ralo xifene, in a dose of 90 mg/day, in suppressing bone turnover and avoiding endometrial and breast stimulation.¹⁹⁸ However, it is being primarily developed in a dose of 60 mg/day given orally for the treatment of vulvar and vaginal atrophy, sites where ospemifene exerts a significant estrogenic impact.¹⁹⁹ There is no effect on vasomotor symptoms. In preclinical studies, ospemifene suppressed the development of breast cancer in the drug-induced mouse model.

Several selective estrogen agonists/antagonists that were first tested for the treatment of breast cancer have potential for the prevention and treatment of osteoporosis. One possibility is droloxifene. Other members of this family with potential use are lasofoxifene and ormeloxifene. Preclinical studies have indicated the possibility of greater efficacy and potency in bone effects, but clinical use will depend on the outcomes of clinical trials.

The results from a 5-year, international, randomized trial with lasofoxifene given to postmenopausal women with osteoporosis indicated a reduction in both vertebral and nonvertebral fractures with a dose of 0.5 mg daily.²⁰⁰ However, the reduction in nonvertebral fractures was small and consisted of forearm and wrist fractures; there was no significant reduction in hip fractures. There was a substantial decrease in the risk of estrogen receptorpositive breast cancer, a modest reduction in coronary heart disease and stroke, and no adverse effects on the endometrium. Treatment was associated with a 2-fold increase in venous thrombosis, hot flushing, and leg cramps. A clear-cut superiority over other treatment options was not apparent.

Bazedoxifene belongs to the estrogen agonist-antagonist family of drugs. It has favorable effects on bone and lipids, but does not affect the endometrium or the breast.^201,202 Bazedoxifene in a dose of 20 mg daily decreased the risk of all clinical fractures in a randomized
clinical trial, with a potency comparable to other anti-resorptive agents in postmenopausal women at high risk for fractures.^203,204 In a subgroup of women at higher risk for fractures, bazedoxifene had a reduced risk of nonvertebral fractures (50% reduction with 20 mg), compared with both raloxifene and placebo. The only adverse event that differed with treatment was an increase in venous thrombosis with treatment compared with placebo; there was neither a beneficial nor an adverse effect on coronary heart disease or stroke. Treatment for 2 years had no effect on mammographic breast density.²⁰⁵ The results of this trial indicate that the effect of bazedoxifene on bone should be comparable to that of estrogen and bisphosphonates.

The reduction of nonvertebral fractures with bazedoxifene compared with raloxifene should not be ignored. We have known for some time that even with 8 years of follow-up, raloxifene has no impact on the risk of hip fractures. This is likely because raloxifene is less potent, and thus the hip with a mixture of cortical and trabecular bone is more resistant to raloxifene’s effects, compared with the spinal column that is composed of sensitive, trabecular bone. Bazedoxifene partnered with estrogen is called TSEC (tissue-selective estrogen complex). The idea is to gain the benefits of estrogen (bazedoxifene by itself has little impact on hot flushes), protect the endometrium and possibly the breast, and enhance some actions of estrogen, such as a reduction in fractures.²⁰⁶ Bazedoxifene combined with conjugated estrogens effectively suppresses hot flushing, improves vaginal atrophy and lipids, prevents bone loss, and does not cause breast tenderness.^207,208 and ²⁰⁹ The combination of 20 mg bazedoxifene with conjugated estrogens prevents endometrial hyperplasia and has an extremely high rate of amenorrhea.^210,211 This approach to postmenopausal hormone therapy may eliminate the need for progestational agents.

Tibolone is structurally related to the 19-nortestosterone progestins that are used clinically in oral contraceptives; however, its activity depends on its metabolism. Tibolone {7-α,17α-17-hydroxy-7-methyl-19-norpregn-5(10)-en-20-yn-3-one} is metabolized among human and nonhuman primates into three biologically active metabolites; the 3β-hydroxy (3β-OH) metabolite and the 3(3-hydroxy (3(3-OH) metabolite have estrogen agonist properties, whereas the Δ-4 ketoisomer has progestogenic and androgenic effects.^212,213

Although tibolone itself binds to the estrogen receptor, in vivo the activity of the 3-hydroxy metabolites is 100 times greater, with a greater affinity for the estrogen receptor-alpha than for estrogen receptor-beta.²¹³ The loss of the hydroxyl group at position 3 of the A ring eliminates estrogenic activity in the δ-4 isomer. The δ-4 isomer exerts its androgenic effects primarily in the liver and brain.

The conversion of tibolone into metabolites occurs chiefly in the liver and intestine. The metabolism of the parent compound is rapid and very near total, yielding mainly the 3α-OH and the 3β-OH metabolites in the circulation; the level of the 3α-OH metabolite is 3-fold higher compared with the 3β-OH metabolite.^214,215 Tibolone and the Δ-4 isomer can be detected only at peak levels 2 h after ingestion, and, even then, the levels are very low, at the limit of detection. The half-life of the metabolites that predominate in the circulation (the 3α-OH and 3β-OH metabolites) is approximately 7 to 8 h, and a steady state is attained by day 5.²¹⁶ Eating does not affect the metabolism, and tibolone can be taken at any time of the day.²¹⁴ The pharmacokinetics of tibolone are not affected by impaired renal function. There is a very weak effect of the metabolites on cytochrome P450 enzymes, and no interference is to be expected with coadministered drugs.²¹⁵ After the initial metabolism of tibolone, the products are rapidly sulfated, and more than 75% of the metabolites circulate as the sulfates to be activated by tissue sulfatases.²¹⁵

Tibolone is available in 2 daily doses, 1.25 and 2.5 mg. There is considerable variability (about 30-40%) within and between subjects, but the 1.25 and 2.5-mg doses produce the same bioequivalence as measured by maximum levels and areas under the curve for the 3a-OH and 3(3-OH metabolites.²¹⁶ However, there are differences in clinical responses, which influence the choice of dose.

The metabolism of tibolone is not limited to the liver and intestine. Important effects are explained by specific local tissue metabolism. For example, the A-4 isomer is primarily produced within the endometrium, binds to the progesterone receptor, and protects the endometrium from the agonist effects of the two estrogenic metabolites.^217,218,219 and ²²⁰

Menopausal symptoms provide the main motivation for women to use postmenopausal hormonal therapy. Tibolone must perform well in this category in order for it to be an attractive option for clinicians and patients. Clinical studies have established without question that tibolone exerts an estrogenic beneficial impact on hot flushing and vaginal dryness. Appropriate studies have documented that tibolone in a daily dose of 2.5 mg is as effective as standard postmenopausal hormone regimens in treating hot flushing.^{221,222,223,224,225,226} and ²²⁷ The 1.25-mg dose takes longer to be effective, and this dose also has a higher incidence of persistent flushing.²²⁸ In addition, tibolone effectively treats the side effect of hot flushing associated with gonadotropin-releasing hormone (GnRH) agonist therapy.²²⁹ Fortunately, tibolone treatment provides an estrogenic effect on the vagina. Tibolone, 2.5 mg daily, relieves vaginal dryness and dyspareunia, and in most studies, tibolone is as effective as estrogen treatment.^{223,224,225,226} and ^{227,230,231,232} and ²³³

A decided advantage for tibolone can be found in studies examining sexuality. In prospective, randomized trials comparing tibolone with estrogen or estrogen-progestin therapy, tibolone has been associated with a better sexual response.^{224,227,234,235,236} and ²³⁷ An increase in libido has been reported in studies comparing tibolone with placebo,^233,238 and the response has been greater than that with estrogen therapy, comparable to that associated with androgen treatment.²³⁹ The overall effect has included an increase in sexual interest and sexual performance, specifically fantasies, arousal, and orgasm.

There are two possible mechanisms for tibolone’s effect on sexuality: a direct androgenic effect of the δ-4 isomer and/or an increase in the circulating level of free testosterone. Tibolone is associated with a profound change in the circulating levels of sex hormonebinding globulin, about a 50% decrease.^233,240 This is undoubtedly due to the δ-4 isomer and an androgenic effect on the liver. Tibolone treatment, therefore, produces a decrease in the concentration of total testosterone (bound and unbound) but a substantial increase in the amount of free, unbound testosterone. This hormonal profile is a striking contrast to that associated with estrogen therapy, which increases sex hormone-binding globulin and decreases both total and free testosterone levels. The androgen side effects of acne and hirsutism, however, have not been reported with tibolone treatment.

When women who had been on tibolone for 10 years were compared with a matched group, the treated women were less clumsy, less anxious in response to mild stress, and demonstrated better memory for facts, although there was no difference in memory for events and worse performance on sustained attention and planning.²⁴¹ Overall, tibolone exerts a positive effect on mood that is modest in impact.^240,242 However, this is an area in which it is not easy to achieve consistent effects, a problem often due to the differences in measurement tools and definitions. The study of cognition is difficult because of the need to match treated and control groups for intelligence, age, occupation, education, and mental state (e.g., depression). Because of this difficulty, the literature reporting the effects of hormone therapy on cognition provides an inconsistent picture. This is further complicated by the sensitivity and appropriateness of the assessment tools that are used. This is an area that requires standardization and new approaches for research, not only for tibolone but for all pharmacologic treatments that affect the central nervous system.

Consistent with reports of tibolone’s effects on HDL-cholesterol in postmenopausal women, tibolone-treated monkeys have much lower HDL-cholesterol compared with control
monkeys.^243,244 Although tibolone treatment resulted in lower circulating HDL-cholesterol, coronary artery atherosclerosis extent in monkeys was not significantly different from the control group. Similar results were observed in the carotid arteries.²⁴⁴ That observation prompted the question of whether the HDL-cholesterol reductions noted among the animals treated with tibolone were associated with physiologically meaningful reductions in HDL-cholesterol function. HDL-cholesterol has a critical role in reverse cholesterol transport, the mechanism by which cell cholesterol (i.e., artery wall cholesterol) can be returned through the plasma to the liver for excretion.²⁴⁵ Further, it has been found that cholesterol efflux capacity predicts the severity and extent of coronary artery disease in human patients.²⁴⁶ Postmenopausal monkeys treated with tibolone had no reduction in cholesterol efflux.²⁴⁷ This disassociation between reductions in circulating concentrations of HDLcholesterol and the lack of change in HDL-cholesterol function suggests the likelihood that this may account in large part for the finding that coronary artery atherosclerosis was not increased or decreased in the monkey model.

Short-term clinical studies uniformly document that tibolone treatment, 2.5 mg/day, reduces HDL-cholesterol levels in women by about 20%; however, there is also a reduction in total cholesterol (about 10%) and triglycerides (about 20%) and a slight decrease or no change in LDL-cholesterol levels.^{235,248,249,250,251,252,253} and ²⁵⁴ In women, therefore, tibolone does not increase LDLcholesterol, and the reduction in HDL-cholesterol is less that that recorded in monkeys. In addition, tibolone decreases LDL-cholesterol oxidation and produces a shift away from small dense LDL-cholesterol (which is more atherogenic); both changes would be beneficial.²⁵⁴ The potential harmful effects associated with reductions in HDL-cholesterol are further balanced by tibolone-associated reductions in endothelin and lipoprotein(a), antiischemic effects detected in women with angina, and an improvement in insulin sensitivity.^{253,255,256,257} and ²⁵⁸ In longer term studies, HDL-cholesterol levels did not come back to baseline at the end of 2 years of treatment, but did return to baseline at the end of 3 years.^253,259,260 and ²⁶¹ And other studies have found that the decrease in HDL-cholesterol is statistically insignificant.^262,263

The recognition that reductions in HDL-cholesterol are potentially harmful is based on the important roles for HDL-cholesterol in the mediation of cholesterol movement from lipid-laden cells and inhibition of LDL-cholesterol oxidation. However, the experimental results in the monkey model indicate that reductions in HDL-cholesterol concentrations are not directly paralleled by reductions in important HDL-cholesterol functions. At least one reason for this lack of direct correlation is the complex nature of HDL-cholesterol lipoproteins, a heterogenous collection of particles that differ in their activities.²⁶⁴ The overall change in HDL-cholesterol levels will not reflect specific changes in particles that can affect specific biologic activities. Similar to results in the monkey model, a randomized trial in women demonstrated that significant reductions in HDL-cholesterol levels (average 27%) caused by tibolone treatment, 2.5 mg/day, were due to a decrease in one subclass of HDLcholesterol particles, and measures of HDL-cholesterol antiatherogenic functions (reverse cholesterol transport and inhibition of LDL-cholesterol oxidation) were not impaired.²⁵⁸ The study was limited by the short, 12-week duration of treatment; however, the findings are consistent with those obtained in the 2-year monkey experiment. These human results were confirmed and strengthened by a study of 68 postmenopausal women randomized to daily treatment for 3 months with either 2.5 mg tibolone or placebo.²⁶⁵ Changes in HDLcholesterol were associated with an increase in hepatic lipase activity, an androgenic effect, again without impairing the ability of plasma to maintain cholesterol efflux.

Results in the monkey model are consistent with an overall neutral impact on the cardiovascular system.²⁴⁴ A long-term (average of 7.5 years) follow-up of women treated with tibolone found no increase in carotid artery intimal media thickness and the number of atherosclerotic plaques, results that are consistent with the monkey model.²⁶⁶ This neutral impact is further supported by failing to find any effect of tibolone on experimentally induced brachial artery dilation or on vascular resistance measured in the carotid and middle cerebral arteries.^262,267 On the other hand, a method studying venous dilation in the hand found an improvement in
endothelium-dependent responses after tibolone treatment.²⁶⁸ Myocardial infarction and heart failure have been reported to be associated with overactivity of the sympathetic component of the cardiac autonomic nervous system, and tibolone treatment decreases plasma levels of free fatty acids, an effect that results in an improved ratio of cardiac sympathetic tone to parasympathetic tone.²⁶⁹ Another favorable effect connected to tibolone and its metabolites is a direct impact on endothelial cells that results in a beneficial decrease in endothelial-leukocyte adhesion molecules, another human finding similar to that in the monkey trial.²⁷⁰

The OPAL study (Osteoporosis Prevention and Arterial effects of tiboLone) was a 3-year, randomized, double-blind trial in six U.S. centers and five European centers, treating 866 postmenopausal women with either 2.5 mg tibolone daily, 0.625/2.5 mg daily of conjugated estrogens/medroxyprogesterone acetate, or placebo.²⁷¹ The arterial endpoint of the study was carotid intima-media thickness measured by ultrasonography every 6 months. Both the tibolone-treated group and the estrogen/progestin-treated group demonstrated an increase in intima media thickness over the time period of the study, at a rate significantly greater than the placebo group, leading to the conclusion that both tibolone and estrogen/progestin treatment increased atherosclerosis compared with the placebo group.

In the OPAL trial, European women differed from American women in multiple ways: higher lipids, higher blood pressure levels, more smokers. Hysterectomized women were excluded in the U.S., but not in Europe (28% of the study population). The overall mean results, indeed, indicated a difference comparing both treatment groups to placebo. But in the European women, atherosclerosis, measured by intima-media thickness, improved in the placebo group, making it easy to calculate a significant difference compared to the treated groups! In American women, there were no differences comparing the three treatment groups, all demonstrated progression of thickness. Thus the overall conclusion was inordinately influenced by the results in the European women. The investigators could not explain these differences. Unfortunately, the OPAL trial did not achieve its goal of providing robust data on cardiovascular effects, due to the older age of the women and the notably different results in American and European women. There continues to be good reason to believe that tibolone has a neutral effect on the cardiovascular system. In addition, tibolone does not adversely affect the blood pressure in women with established hypertension.²⁵¹

A case-control study assessed the risk of venous thromboembolism in a very large population of postmenopausal women (23,505 cases and 231,562 controls) derived from the U.K. General Practice Research Database.²⁷² No increase in risk was observed with the current use of either tibolone or transdermal estrogen compared with a significant increase associated with the current use of oral estrogen.

The administration of tibolone, 2.5 mg/day, to older women with type 2 diabetes mellitus produced no significant changes in the lipid profile.²⁷³ Tibolone treatment is associated with an increase in insulin sensitivity in women with insulin resistance, although some have reported no effect in normal women.^{250,258,274,275} Therefore, tibolone is an attractive option for postmenopausal women with diabetes mellitus.

Tibolone does not stimulate endometrial proliferation. This is because the predominant if not exclusive tibolone metabolite produced within the endometrium is the Δ-4 isomer,
which binds to the progesterone receptor and protects the endometrium from the agonist effects of the two estrogenic metabolites.^217,218,219 and ²²⁰ This protective effect has been documented in long-term (up to 8 years) human studies.^{219,220,223,231,232,276,277} and ²⁷⁸ Isolated cases of endometrial proliferation have been reported, for example, 4 of 150 women treated with 2.5 mg daily for 2 years.²⁷⁹ In a 5-year follow-up, 47 of 434 women experienced bleeding, and of these 11 had endometrial polyps or fibroids, but there were two with simple hyperplasia and two with endometrial carcinoma in situ.²⁸⁰ This underscores the standard clinical principle to investigate persistent vaginal bleeding in any postmenopausal woman. In the major U.S. clinical trial, three cases of endometrial cancer were observed, but, in each case, pre-existing carcinoma was later detected when the initial biopsy samples were more extensively examined.²⁵³ Nevertheless, a second large 2-year trial was conducted, the THEBES Study, and no endometrial hyperplasia or cancer occurred in the tibolonetreated groups.²⁸¹

The reported breakthrough bleeding rates with tibolone treatment have been comparable to treatment with combined, continuous estrogen-progestin therapy^224,249,276 and ²⁷⁷; but welldesigned comparison clinical trials indicate that the rate is less with tibolone.^{225,227,233,282,283} In addition, amenorrhea is achieved more rapidly; 90% of tibolone-treated women are amenorrheic by 6 months.^225,280,284 Bleeding is less in older women and can be greater with the 2.5-mg dose compared with the 1.25-mg dose but the difference is too small to be detected in some studies.^228,253,285 Importantly, a lack of correlation has been observed between bleeding and endometrial thickness measured by ultrasonography.^285,286 This again emphasizes the need to biopsy tibolone-treated women with persistent bleeding.

Careful evaluations of women with fibroids who have been treated with tibolone have revealed no evidence of myoma growth, with up to 3 years of follow-up.^287,288 and ²⁸⁹ Furthermore, add-back treatment with tibolone effectively prevents flushing and bone loss and does not impair the fibroid response to therapy with GnRH analogs.²⁹⁰

The breast is a complicated estrogen factory. Breast tissue, normal and abnormal, contains all the enzymes necessary for the formation of estrogens (sulfatase, aromatase, and 17β-hydroxysteroid dehydrogenase) and the conversion of estrogens into their sulfates (sulfotransferase). Estrone sulfate concentrations are high in the breast (higher than in plasma) and even higher in cancer tissue. This state is achieved in postmenopausal women with very low systemic levels of estrogen, indicating that a local mechanism is operative.

The major pathway of estrogen synthesis in human breast tumor cells is by conversion of estrone sulfate to estrone by estrone sulfatase, a pathway that is more important that the aromatase pathway.²⁹¹ Aromatase is an enzyme complex that produces the irreversible conversion of androgens to estrogens. The location of aromatase activity is predominantly in the stromal tissue of the breast. Comparing normal to tumor tissue, the levels of estrone sulfate and estradiol were higher in the tumor tissue.²⁹² Sulfatase activity is higher (130-200 times) than aromatase activity in all breast tissues examined, and the sulfatase and aromatase activity was higher in the tumor tissue than in normal tissue. Thus, estrogen concentrations in the breast are higher in women with breast cancer, and formation of estradiol from sulfated estrogen is the primary pathway. Most importantly, this increase in estrogen activity is independent of the estrogen receptor status of the tissue.

Tibolone and its metabolites inhibit estrone sulfatase and 17β-hydroxysteroid dehydrogenase in normal stromal cells and in hormone-dependent breast cancer cells (MCF-7 and T-47D).^293,294,295 and ²⁹⁶ This inhibits conversion of estrone sulfate to estradiol. In addition, tibolone
and its 3-hydroxymetabolites increase the conversion of estrone back to estrone sulfate by increasing the activity of sulfotransferase.²⁹⁷ Tibolone and all three metabolites inhibit the conversion of estrone to estradiol by 17β-hydroxysteroid dehydrogenase.²⁹⁴ Although these effects resemble progestin activity, tibolone is more potent. Tibolone increases aromatase activity in stromal cells but only at high concentrations that are beyond in vivo levels.²⁹⁵ These tibolone-induced enzyme changes would lower the active estrogen concentrations in breast tissue.

In the rat and mouse breast cancer models (cancer induced by 7,12-dimethylbena{a} anthracene, DMBA), tibolone exerts protective effects to the same degree as tamoxifen.²⁹⁸ However, tibolone is not antiestrogenic and does not inhibit aromatase. Therefore, the mechanism is explained by the enzyme effects summarized previously, inhibition of sulfatase and 17β-hydroxysteroid dehydrogenase and stimulation of sulfotransferase to increase the production of inactive sulfates.²⁹⁵ In addition, tibolone increases cellular differentiation and stimulates apoptosis, at least with normal breast cells in vitro.²⁹⁹ An increase in apoptosis is an action of the parent tibolone and its δ-4 isomer. Thus, tibolone acts like progestins and weak androgens in breast cell line studies examining proliferation, differentiation, and apoptosis.

Tibolone and its metabolites do not display the same activity directed toward the sulfatase enzyme in all tissues. Strong inhibition of sulfatase is a major feature in breast cells, but tibolone and its metabolites inhibit sulfatase only moderately in the endometrium (contributing to an antiestrogenic action) and provide no inhibition in bone (allowing a greater estrogenic impact).³⁰⁰

Postmenopausal hormone therapy increases breast density on mammography in about 10-20% of estrogen users and about 20-35% of estrogen-progestin users, an effect that occurs within the first months of treatment. In contrast, tibolone does not increase breast density and causes far less mastalgia than that seen with estrogen treatment.^{225,237,253,283,301,302,303,304} and ³⁰⁵ It is logical to conclude that these favorable responses are a consequence of the tibolone effects on the breast tissue enzymes involved in local estrogen production.

The Livial Intervention following Breast cancer: Efficacy, Recurrence, And Tolerability Endpoints (LIBERATE) trial was a multinational, placebo-controlled, randomized study of women with vasomotor symptoms who had had breast cancer surgically treated within
the previous 5 years.³⁰⁶ The study was designed to demonstrate that tibolone was superior to placebo, but when the drug monitoring safety board notified the sponsor that there appeared to be an excess of breast cancers in the treated group, the trial was canceled July 31, 2007, 5 months before its scheduled end. The median duration of participation and treatment was about 3 years, with a wide range from a few weeks to almost 5 years. The participants used a variety of adjuvant treatments, mostly tamoxifen, 66.8%; 6.5% used aromatase inhibitors. The dose of tibolone was 2.5 mg daily. Final numbers for analysis were 1,556 women in the treated group and 1,542 in the placebo group. The women ranged in age from under 40 to 79, with a mean age of 52.7 years. 57.8% had positive lymph nodes and 70% had a tumor stage of IIA or higher. Estrogen receptor status was known in 2,808 women in whom the tumors were estrogen receptor positive in 77.8%. In the intent-to-treat analysis, the hazard ratio for recurrent breast cancer in the tibolone-treated women was 1.40 (CI=1.14-1.70). The absolute risk for tibolone was 51 cancers per 1,000 women per year, and 36 in the placebo group. The increase occurred only in women with estrogen receptor-positive tumors. There was no difference in mortality rates between the two groups during the 5-year study period. There were no differences in cardiovascular events or gynecologic cancers, and not surprisingly, vasomotor symptoms, quality of life measures, and bone density improved with tibolone treatment.

How do the LIBERATE results that indicate an estrogenic action of tibolone in breast cancer survivors jibe with the literature indicating that tibolone exerts a non-estrogenic effect on breast tissue? Indeed, it was realistic to expect tibolone to have a salutary effect on the breast. It is well documented that the breast responds to tibolone with less stimulation compared with estrogen, judged by changes in mammographic breast density and the characteristics of tissue obtained by fine-needle aspiration. In the LIFT clinical trial (discussed under bone) that had vertebral fractures as the primary endpoint and breast cancer as a secondary endpoint, the risk of breast cancer after 3 years was significantly 68% reduced with tibolone treatment, although the dose was lower, 1.25 mg daily.³⁰⁷

The previous literature documenting beneficial actions of tibolone on the breast reflected, however, the impact of tibolone on normal breast tissue, and tibolone’s activity to lower local bioactive estrogen levels in target tissues might be lost in cancer cells. The contrary results in the LIFT trial could reflect its older population of women at high risk for fractures, a population that also differed by having lower body weights, no history of tamoxifen treatment, and lower risk factors for breast cancer.

Although the LIBERATE trial may apply to all breast cancer survivors, speaking strictly in a scientific sense, the results were derived mainly from tamoxifen users with 10-fold fewer users of aromatase inhibitors. The possibility that estrogen or tibolone would interfere with the beneficial effects of tamoxifen or aromatase inhibitors has always been one of the objections to treating breast cancer survivors with estrogenic hormones. In a subgroup analysis of the LIBERATE trial, the group of women who had used aromatase inhibitors had a greater risk of recurrent breast cancer compared with tamoxifen; however, the confidence interval was wide because of relatively small numbers. Possibly the estrogenic effect of tibolone would be more pronounced on an occult breast cancer in estrogen-depleted tissue compared with tissue where tamoxifen was bound to the estrogen receptor and prevented estrogenic stimulation. We don’t know if the LIBERATE data are meaningful for future treatment regimens. Nevertheless, until there are new data, the use of tibolone in women with a history of breast cancer remains relatively contraindicated.

Tibolone prevents bone loss in postmenopausal women as effectively as estrogen or estrogen-progestin therapy.^{235,253,308,309,310,311} and ³¹² The beneficial impact on bone can be attributed to the
estrogenic metabolites acting through the estrogen receptor because it is blocked by an antiestrogen but not by an antiandrogen or an antiprogestin.³¹³ In a large, U.S. dose-response study with doses ranging from 0.3 to 2.5 mg daily, only the 1.25 and 2.5-mg doses produced progressive bone density increases in the femoral neck.²⁵³ Indeed, the impact on bone was essentially the same for the 2 highest doses, 1.25 and 2.5 mg. Although the 1.25-mg dose is acceptable for the prevention of bone loss, the 2.5-mg dose is more effective for the alleviation of hot flushes.²²⁸ Tibolone prevents the bone loss associated with GnRH agonist treatment (and the side effect of hot flushing).^290,314

The LIFT study (Long-term Intervention on Fractures with Tibolone) was a randomized, placebo-controlled multicenter trial in 22 countries of tibolone, 1.25 mg, given daily over 3 years.³⁰⁷ The 4,538 women who participated in the trial were age 60 to 85, all at high risk of fractures because of osteoporosis, and all treated with calcium and vitamin D supplementation. The study was stopped in February 2006 after a mean treatment of 34 months because of an increased risk of stroke. The risks of all events were assessed after 5 years of follow-up. Tibolone treatment reduced the number of vertebral fractures by about 45%, and nonvertebral fractures by about 25%. The reduction of fractures was about 4 times as great in women who already had a vertebral fracture upon entry to the study compared with women who had not had a fracture at baseline. It is noteworthy that the number of falls in the treated group was 25% less.

Based on previous bone density studies, the results of the LIFT trial on fracture reduction were not unexpected. The magnitude of the effect is roughly comparable to those with estrogen, bisphosphonates, and raloxifene (with the important exception being a lack of effect of raloxifene on hip fractures). The reduction of breast cancer was comparable to that reported with tamoxifen and raloxifene, but this was not a primary endpoint of the study. Although the difference was not statistically significant, there were four cases of endometrial cancer in the tibolone group and none in the placebo group.

The reported 2-fold increase in stroke in the LIFT trial was greater in the oldest women (over age 70), similar to that observed with estrogen. It is best to avoid the use of tibolone in elderly women and in women who are at risk for stroke (specifically those with hypertension, smoking, diabetes, or atrial fibrillation).

The treatment of choice for vasomotor symptoms is hormone therapy. However there exists a substantial number of women who either cannot or will not accept hormone therapy. The choices other than hormone therapy that have been available in the past offered only a modest benefit. Transdermal clonidine, applied with the 100-μg dose once weekly was a common choice, but the reduction in hot flushing was only slightly better than that obtained with placebo treatment.^317,318 Clonidine, bromocriptine, and naloxone given orally are only partially effective for the relief of hot flushes and require high doses with a high rate of side effects such as drowsiness and dry mouth. Bellergal treatment (a combination of belladonna alkaloids, ergotamine tartrate, and phenobarbital) is slightly better than a placebo and a potent sedative in the short-term; however, one study documented a similar response with bellergal and placebo after 8 weeks.^319,320 Veralipride, a dopamine antagonist that is active in the hypothalamus, is relatively effective in inhibiting flushing at a dose of 100 mg daily, but is associated with major side effects, including mastodynia and galactorrhea.^321,322 and ³²³ Medroxyprogesterone acetate (10-20 mg daily) and megestrol acetate (20 mg daily) are also effective (as effective as estrogen), but concerns regarding exogenous steroids, especially in patients who have had breast cancer, would apply to progestins as well.^324,325 and ³²⁶ Vitamin E, 800 IU daily, is barely more effective than placebo.³²⁷ Dong quai, ginseng, black cohosh, isoflavones (including soy protein), yoga, and acupuncture all have little clinical difference compared with placebo treatment.^{328,329,330,331,332,333,334,335,336,337} and ³³⁸

In the last few years, selective serotonin reuptake inhibitors (SSRIs) have gained a reputation for significant efficacy in treating hot flushing. The drugs that have been studied include citalopram (Celexa), fluoxetine (Prozac), sertraline (Zoloft), paroxetine (Paxil), and serotonin and norepinephrine reuptake inhibitors, venlafaxine (Effexor) and desvenlafaxine succinate (Pristiq). In addition, an antiseizure medication, gabapentin (Neurontin), has been demonstrated to reduce vasomotor symptoms.

In the study with paroxetine (the controlled release product), 61% of the treated group (a general population of postmenopausal women with only 12 individuals who were breast cancer survivors) at the end of the study had at least a 50% reduction in frequency and severity of flushing, an effect that was about 2.5 times better than placebo with the higher dose.³³⁹ Venlafaxine was studied in breast cancer survivors; although the optimal dose was 75 mg, an appreciable response with 37.5 mg indicated that it would be worthwhile to begin treatment with the lower dose.^340,341 The response was very rapid, within days, and therefore the dose can be increased in 1-2 weeks. The main side effects were mouth dryness, anorexia, nausea, and constipation. The efficacy of venlafaxine was demonstrated to be the same in women taking or not taking tamoxifen. Desvenlafaxine succinate is a metabolite of venlafaxine and equally effective as the parent compound.^342,343 The effects of citalopram, fluoxetine and sertraline are no more effective than placebo, about a 50% reduction in short-term trials.^{344,345,346,347} and ³⁴⁸

Gabapentin (Neurontin) is a γ-aminobutyric acid analogue that has been used for seizures since 1994. It is also effective for migraine headaches, tremors, and panic disorder. In a gabapentin clinical trial, 67% of the treated women experienced more than a 50% reduction in flushes at week 12, compared with 38% in the placebo group.³⁴⁹ The most common side effects were somnolence (20%) and dizziness (13%). Peripheral edema occurs occasionally because of an induced decrease in serum protein. The potency of this agent appears to be more modest than the SSRIs in a dose of 900 mg/day.^350,351 At higher doses, gabapentin was as effective as estrogen (about a 70% reduction in flushing); however, side effects are common at higher doses.³⁵²

Pregabalin (Lyrica), a more potent form of gabapentin, is an anticonvulsant drug that has been used in doses of 150-300 mg daily to treat anxiety and neuropathic pain, e.g., diabetic neurogenic pain, pain after shingles, and fibromyalgia. Side effects include dizziness, drowsiness, visual disturbances, tremor, weight gain, and a decrease in libido. In a phase III randomized trial, a dose of 75 mg b.i.d. (the recommended dose because of more side effects with higher doses), hot flushing was reduced by 65% after 6 weeks of treatment, an impact that was only 15% greater than placebo.³⁵³ Although the data are limited to this short-term, small clinical trial, the effect of pregabalin appears to be comparable to gabapentin.

Tamoxifen is converted to an active metabolite by enzymes that are inhibited by certain SSRIs. Paroxetine coadministration decreases plasma concentrations of the active tamoxifen metabolite.^355,356 A lesser effect is associated with fluoxetine and sertraline. In a retrospective cohort study, only paroxetine use during tamoxifen therapy was associated with an increased risk of death due to breast cancer.³⁵⁷ Paroxetine, fluoxetine, and sertraline are best avoided in women being treated with tamoxifen.

The operations of a pharmacy are regulated in the individual states by state boards of pharmacy, in a system similar to the regulation of medical practice. The first federal law regulating drugs, the Federal Food, Drug and Cosmetic Act, was passed in 1938, at a time when most drugs were compounded according to a doctor’s prescription. The American Pharmaceutical Association defines pharmacy compounding as the preparation of a prescription drug that is “individualized” to the needs of the patient. This changed after World War II with the development and growth of the pharmaceutical industry. The Kefauver-Harris Amendment in 1964 extended the role of the Food and Drug Administration (FDA) to include safety and efficacy.

In the 1990s, the FDA began to regard the drugs coming from compounding pharmacies as falling under the “new drug” regulations, and therefore, the FDA had jurisdiction over the marketing and promotion of those drugs. The pharmacy world was immediately challenged; there was no way that an individual pharmacy could carry out the kind of clinical studies required for the approval of new drugs. Thus, the pharmacists immediately realized that all compounded drugs would be illegal. At the same time, the FDA was a bit ambivalent, acknowledging that there were examples where the individual needs of a patient required the compounding of a drug, e.g., the creation of a liquid preparation when none was available. This was before compounding took to the Internet for marketing and promotion.

In 1992, the FDA issued its Compliance Policy Guide on Compounding, reserving a right for “selective enforcement,” as a compromise between believing it was correct in assuming that compounded drugs represented new drugs and admitting that some patients required compounding. The pharmacy profession immediately rejected the idea that the FDA had any regulatory jurisdiction over pharmacies. The 1997 Food and Drug Modernization Act attempted to clarify the situation. An amendment was added to the existing laws stating that compounded drugs were not “new drugs,” but at the same time the 1997 act prohibited the marketing of compounded drugs.

The pharmacy profession sued the FDA, arguing that a restriction on advertising and promotion of compounded drugs was unconstitutional, a restriction of free speech. The District Court ruled against the FDA in 1999. The FDA appealed and lost again in the 9th Circuit Court of Appeals, which further invalidated the entire 1997 act. In 2002, the U.S. Supreme Court upheld the Circuit Court decision.

The FDA issued a new Compliance Policy Guide in 2002, stating that selective enforcement would hinge on 3 major factors: (1). A potential adverse effect of a drug, (2). Whether drugs were compounded from non-FDA-approved components, and (3). Whether compounded drugs were similar to drugs already removed from the market for safety reasons. At this point, the FDA affirmed that it did not want to infringe on the traditional practice of compounding, the preparation of a drug according to a doctor’s prescription to fit an individual patient’s requirements.

Wyeth Pharmaceuticals filed a Citizen Petition with the FDA in October 2005 requesting that the FDA take action against several compounding pharmacies that were primarily
Internet-based. The Petition’s major allegation was that these pharmacies were essentially manufacturing new drugs and should be subject to new drug regulations. On January 9, 2008, the FDA announced it would take action against seven pharmacies providing prescription bioidentical hormones, and issued warning letters that potentially could be followed by seizures of drugs and injunctions against production.

The FDA would like to regard compounded drugs as “new drugs,” but the legal precedent has now been set by the courts: compounded drugs are not “new drugs.” This was reaffirmed in a 2006 decision in the U.S. District Court for Texas. The FDA would further like to regard the giant compounding pharmacies, especially those operating over the Internet, as manufacturers, like pharmaceutical companies, but again, the court decisions have prevented the FDA from requiring compounding pharmacies to meet “new drug” standards. The “new drug” argument doesn’t work.

“Bioidentical” and “natural” are often used in concert. Strictly defined, the hormones must be precisely the same as normal endogenous estrone, estradiol, and estriol, the three endogenous estrogens; progesterone, the progestational agent synthesized by the ovarian corpus luteum after ovulation, and testosterone and dehydroepiandrosterone, androgens made by the human body. To argue that products are not “artificial,” begs the issue because even if the source is a steroid molecule derived from plants, chemical and manufacturing processing is still required. These terms obviously have marketing value, and the terms are used to imply greater safety, even greater efficacy. The situation is further compounded (pun intended) because it is likely most patients assume that the marketed bioidentical and natural hormones have been demonstrated by appropriate studies to be effective and safe. Of course this is not the case, although it seems like an obvious conclusion to view product A and product B to be the same if they are the same molecule. The problem is that compounding pharmacies are not required to compare the formulation with the performance of an approved product, nor is there any way for a patient to be assured that the dosage is correct (that the drug contains what it is supposed to contain).

The large compounding pharmacies meet the “individualization” requirement that usually comes from a clinician-patient interaction, by promoting salivary measurements of hormones, as interpreted by one of their employed clinicians to produce tailored hormone choices and doses. Assessment of this approach by researchers, as well as organizations such as the American College of Obstetricians and Gynecologists and the Endocrine Society, have concluded that the variations in salivary sex steroid levels from individual to individual and from specimen to specimen preclude clinical interpretation.^359,360 and ³⁶¹ For most patients, laboratory testing is not necessary in hormone decision-making.

The American Pharmacists Association and the National Association of Boards of Pharmacies define compounding as the steps required in order to provide a drug in response to a clinician’s prescription according to an individual patient’s needs, and the preparation of drugs in anticipation of a demand. Therefore, there are three people involved: the patient, the clinician, and the pharmacist. This is in contrast to the production of large amounts of a drug for a national market of unknown users. The American Pharmacists Association further says that if an FDA-approved product is commercially available that meets a patient’s needs, it should be the drug provided.

The key to the position of the pharmacists is the contention that there are circumstances, decided by the patient that makes the use of commercial products not a good choice. This seems reasonable, but it is also reasonable that this decision requires the involvement of the
clinician because ultimately a prescription is still required. The traditional view of compounding, therefore, is one of personal relationships with patient, clinician, and pharmacists. This becomes a totally different story when a large Internet pharmacy responds to thousands of prescriptions with no knowledge of the patients. What happened to the “individualization” aspect of compounding? Is it practicing medicine by the pharmacy to have a registered clinician employed by the pharmacy to interpret hormone levels and adjust doses?

Clinicians are appropriately frustrated by the claims made that bioidentical compounded drugs have greater efficacy and safety. Bruce Patsner argues that it should be accepted that bioidentical drugs from the big compounding pharmacies do not meet the definition of compounding supported by the pharmacist’s own organization, the American Pharmacists Association—a personal relationship of patient, clinician, and pharmacist addressing an individual’s needs.³⁵⁸ The FDA can argue that the big operations are not legitimate compounding, but big commercial operations directed to unknown consumers.

Patsner also argues that the most vulnerable point is the false safety and efficacy claims. The contention should not be that the safety and efficacy claims are inaccurate, because the pharmacies can always compose their words to avoid legal assaults. The point of emphasis should go back to the pharmacist’s published credo: if a commercial, approved product is available to meet the patient’s needs, a compounded product is not indicated. Replacing a commercial product with a “natural,” untested, unregulated product is not the same thing as prescribing a compounded product when no commercial product will meet the individual’s needs.

Pastner summarizes his argument by saying that the large compounding pharmacies are not true compounders because they advertise and promote their products as replacements for commercially available, approved and tested drugs, and that the attempt at “individualization” uses an unsubstantiated method that marginalizes clinicians.³⁵⁸

The bioidentical hormones and the various commercial female hormone products are produced by pharmaceutical companies using similar methods that start from the same raw material, usually soy or yams. Some of the commercial products available consist of estradiol, testosterone, and progesterone, the exact same steroid drugs provided by compounding pharmacies. A major difference between the commercial products and products from compounding pharmacies is the important fact that commercial products are federally regulated and tested for purity and potency; compounding pharmacies have not been regulated in this fashion.

The compounded estrogen formulations that contain combinations of estrone, estriol, and estradiol contain sufficient estradiol to produce the same biologic effects associated with commercial preparations. There are no clinical studies documenting that these combinations confer better results or safety. Whether the presence of estriol reduces the risk of breast cancer has not been tested in appropriate clinical studies. Case-control data indicate that estriol used without a progestational agent increases the risk of endometrial cancer, thus its biologic behavior is similar to that of other estrogens.³⁶²

Salivary sex steroid levels vary widely between individuals, and from measurement to measurement within the same individual. Most importantly, the appropriate clinical studies to document correlation of salivary hormone levels with clinical state or responses have not been performed. Tailor-making a hormone therapy regimen according to
salivary testing is an appealing idea that has not been addressed scientifically, and given the variability in salivary hormone levels, it is unlikely that clinical studies would yield useful information.

“Phytoestrogens” is a descriptive term applied to nonsteroidal compounds that have estrogenic activity or are metabolized into compounds with estrogen activity. Phytoestrogens are classified into three groups: isoflavones, lignans, and coumestans.^364,365 They are present in about 300 plants, especially legumes, and bind to the estrogen receptor. Soybeans, a rich source of phytoestrogens, contain isoflavones, the most common form of phytoestrogens, mainly genistein and daidzein, and a little glycitin.

Isoflavones exist in plants bound as glycoside conjugates attached at the 3 position, called glycones. The carbohydrate component requires gut bacteria to remove the sugar moiety to produce active compounds, the aglycones. Individual variability in gastrointestinal microflora, as well as absorption, influences the bioavailability of isoflavones. Biochanin and formononetin are methylated precursors that are metabolized to genistein and daidzein. Red clover and lentils contain significant amounts of these precursors. The isoflavones are in the active, deconjugated forms in fermented soy foods like miso and tempeh. The concentration of isoflavones in tofu is highly variable.

The phytoestrogens are characterized by mixed estrogenic and antiestrogenic actions, depending on the target tissue and local estrogen concentration. Variations in activity may also be due to the fact that the soy phytoestrogens have a greater affinity for the estrogen receptor-β compared with estrogen receptor-α, although the affinity for the beta receptor is still only 35% that of estradiol.³⁶⁶ Despite a low affinity for the alpha receptor, circulating levels many times that of steroidal estrogens produce the potential for biologic activity.

You can eat soybeans every day and never see a bean. Soybeans are defatted to produce soy flour. Soy flour is then prepared to remove the carbohydrates. 95% of soy flour is toasted and used as animal feed. Alcohol washing is used to get a taste-free product, but alcohol extraction removes the phytoestrogens.³⁶⁷ SUPRO, known as “isolated soy protein,” from Protein Technologies International, the major supplier for commercial products and research, is extracted by aqueous washing and retains the isoflavones.

The reason why most of the soybean crop is devoted to animal feed is because what is left after removing lipids is totally bland. The solution is to mix soybeans with other foods, e.g., beans and soups. Unfortunately they require standing in water for about 12 h and simmering for 2-3 h to be cooked. The average Japanese intake of isoflavones is about 50 mg/day.³⁶⁸ The rest of Asia has an average consumption of about 25-45 mg/day, and Western consumption is less than 5 mg/day.^369,370

A belief that Asian women report fewer menopausal symptoms has been an underlying force in the promotion of isoflavones. However, this apparent difference in the prevalence of symptoms comparing Asia and the West may reflect cultural differences and not actual experience. An Italian study found a 45% reduction in flushing with 60 g of isolated soy protein daily (76 mg isoflavones), compared with a 30% reduction in the placebo group.³⁷¹ Two other studies, both with 50 mg/day of isoflavones, found a similar 15% reduction in the number of flushes compared with placebo.^372,373 Another placebo-controlled short-term trial found a greater reduction in flushes with 70 mg isoflavones daily.³⁷⁴ In a randomized, crossover study of a high dose of isoflavones, 150 mg/day, for flushes in breast cancer survivors, the treated group and the placebo group demonstrated equal effects.³³¹ The dose of 150 mg isoflavones per day was similar to three glasses of soy milk daily. Two Italian randomized trials found the same response to placebo and 72 or 80 mg/day isoflavones.^{375, 376} An Australian study randomized women to 118 mg/day isoflavones or placebo and could detect no difference after 3 months in hot flushing, libido, vaginal dryness, or any of a long list of symptoms, and a Finnish randomized trial using 114 mg isoflavones found no effects on the vagina or on menopausal symptoms.^377,378 In a randomized study in Iowa, no differences were found in hot flush frequency comparing isoflavone-rich soy protein to a whey protein control.³³² And finally, another randomized trial of breast cancer survivors found no difference comparing placebo with 90 mg isoflavones daily.³³³

Ginkgo biloba is an extract prepared from the leaves of the G. biloba tree. It contains flavonoids and unique terpene lactones. Ginkgo biloba is a multimillion dollar herb sold in the U.S. for the preservation of memory. In vitro studies suggested that ginkgo had antioxidant (from the flavonoids) and anti-amyloid (from the lactones) effects. Indeed, the biologic studies provided a rationale for the use of ginkgo to prevent dementia.

A randomized, double-blind, placebo-controlled trial comparing Ginkgo biloba with placebo for the prevention of dementia enrolled 3,069 elderly individuals (over age 75) in five academic centers in the U.S.^395,396 The participants were randomized to b.i.d. doses of 120 mg ginkgo or placebo (45% female in the ginkgo group and 47% female in the placebo group). The ginkgo formulation and dosage were that used in many of the brands sold in the U.S. The dementia rate steadily increased in both groups over a 7-year period of follow-up, accumulating 277 cases (17.9%) in the treatment group and 246 cases (16.1%) in the placebo group. The rate of dementia did not differ between the two groups, nor did the rate of Alzheimer’s disease. In addition, treatment with Ginkgo biloba did not produce less cognitive decline in either adults with normal cognition or with mild cognitive impairment. Other randomized trials have failed to demonstrate any beneficial effects on Alzheimer’s, learning, memory, attention, verbal fluency, or concentration.^397,398

The American trial robustly demonstrated that Ginkgo biloba in the tested and commonly used dose did not delay the onset of dementia or cognitive decline.^395,396 The concept of “delay” is important. A treatment that could delay the onset of dementia by 5 years would reduce the number of dementia cases by 50%. In fact, this clinical trial found a statistically significant increase in the risk for developing dementia with ginkgo treatment in the 25% of participants who had cardiovascular disease prior to enrollment. However, the authors appropriately urged caution in interpreting this subgroup analysis. A Cochrane review in 2007 of 35 clinical trials with 4,247 participants concluded that there was no convincing evidence that ginkgo treatment benefited individuals who already had dementia or cognitive impairment.³⁹⁹

St. John’s wort has been reported to be comparable to tricyclic antidepressants in treating mild to moderate depression, based on eight appropriate trials.⁴⁰⁰ This is the conclusion of two meta-analyses.^401,402 All studies were short-term, about 4-6 weeks in duration, and with small numbers. The treatment consisted of a 300-mg plant extract in tablet form, administered t.i.d. However, 2 large, American 8-week trials found no difference between treatment and placebo.^403,404

The U.S. Food and Drug Administration (FDA) issued an alert in February 2000 that St. John’s wort may interact with drugs known to be metabolized by the cytochrome P450 pathway: theophylline, digoxin, immune suppressants, and oral contraceptives.⁴⁰⁵ St. John’s wort activates an orphan receptor that induces the expression of metabolic enzymes.⁴⁰⁶ In clinically depressed individuals being treated with prescription antidepressants, manic reactions can result (the central serotonergic syndrome).

The cardiovascular story with phytoestrogens received a large boost in 1995, when a metaanalysis concluded that an intake of an average of 47 g soy protein/day lowered total cholesterol and LDL-cholesterol.⁴⁰⁷ This was supported by studies in the monkey indicating that isoflavone increased HDL-cholesterol, enhanced vasodilation, and decreased atherosclerosis.⁴⁰⁸

Only intact soy protein has a beneficial effect on lipids. Separation of the protein component from dietary soy protein loses the effect. This effect depends on the inhibition of cholesterol absorption by the non-isoflavone protein.^409,410 The mechanism involves upregulation of the LDL-cholesterol receptor and catabolism of LDL-cholesterol, leading
to an increase in bile excretion. The soy peptide binds bile acids and prevents resorption. Alcohol extraction removes the isoflavones from soy protein and causes a loss of the beneficial effect on atherosclerosis in monkeys.⁴¹¹ Thus, both the isoflavone portion and the protein component are required for a full cardiovascular effect. Non-alcohol-washed soy protein extract has been extensively studied in monkeys. This preparation lowers total cholesterol and LDL-cholesterol, and raises HDL-cholesterol,^412,413 produces coronary artery vasodilation,⁴¹⁴ inhibits reduction in coronary flow after collagen induced platelet aggregation and serotonin release,⁴¹⁵ and inhibits atherosclerosis but not as robustly as estrogen.^408,413

In women, soy protein reduces total and LDL-cholesterol and does not affect triglycerides or HDL-cholesterol; ethanol-extracted soy protein has no effect.^416,417,418 and ⁴¹⁹ The minimal dose is about 60 mg isoflavones daily, which is present in 25 g soy protein/day.⁴²⁰ LDL-cholesterol must be above 130 mg/dL in order to have an effect. Studies of healthy men and women could detect no effect of phytoestrogens (25 to 80 mg isoflavones per day) on lipids or brachial vasodilation.^421,422 and ⁴²³ In a 12-week study of women with type 2 diabetes mellitus, dietary supplementation of 30 g soy protein (132 mg isoflavones) daily improved insulin resistance and glucose control in addition to lowering total cholesterol and LDL-cholesterol levels.⁴²⁴ In addition, soy intake prevents LDL-cholesterol oxidation in hyperlipidemic men and women even when circulating LDL-cholesterol levels are unaffected.⁴²⁵ Promensil, in a 10-week study, had no effect on lipids (it only contains isoflavones, no protein) but did improve arterial compliance.⁴²⁶

The U.S. FDA, in October 1999, authorized the use in food labeling of health claims related to the association between soy protein and reduced risk of coronary heart disease, “based on the totality of publicly available scientific evidence, soy protein included in a diet low in saturated fat and cholesterol may reduce the risk of CHD by lowering blood cholesterol levels.”⁴²⁷

Remember that both protein and isoflavones are needed for a cardiovascular effect. Isoflavones by themselves have no effect on lipids.^373,426,428 Protein without isoflavones has no effect on vasodilation and atherosclerosis.⁴¹¹ The FDA has stated that there is insufficient evidence to allow them to exclude alcohol-washed products from the health claim, but it makes sense that a combined protein-isoflavone product is best. Even in older women with moderate hypercholesterolemia, a high intake of soy phytoestrogens (purified isoflavones without protein) had no effect on the lipid profile.⁴²⁹ And also remember, that there is no effect on the lipids in individuals who already have a normal profile. Even in individuals with high cholesterol levels, the beneficial impact of soy protein intake is modest and likely to have little clinical effect.

It will require appropriate clinical trials to determine how phytoestrogens compare in the cardiovascular system with estrogens and to determine the efficacy, safety, and correct dosage (studies thus far recommend a daily intake of 60 g soy protein). In addition, the intake of sufficient soy to produce a clinical response is not easy; intake is handicapped by gastrointestinal symptoms, a major alteration in diet or the use of an unpalatable supplement, and great variability in plant contents and products (due to processing). A dietary intake to match the isoflavone dose used in the studies on the lipid profile, for example, would require about 1 lb daily of tofu! In addition, individuals demonstrate great variability in absorption and metabolism. A user-friendly preparation must be developed that minimizes individual variability in response.

Phytoestrogens are effective in preventing bone loss in rats but not in monkeys.^430,431 and ⁴³² In women, most studies have demonstrated at best a slight effect on spinal bone but no effect on hip bone.^{30,417,433,434} One 3-year, randomized trial demonstrated no effect on bone loss
in the spine and femur, with perhaps a modest bone-sparing at the femoral neck with 120 mg/day of isoflavones after adjustment for age and body fat.⁴³⁵ A 1-year clinical trial could detect no impact of soy intake on bone mineral density in either equol producers or nonproducers.⁴³⁶ Flaxseed supplementation had no effect on biomarkers of bone metabolism.⁴³⁷ The difference between hip fracture incidence in Japanese and American women may be due to structural and/or genetic differences not dietary intake.⁴³⁸

Ipriflavone is a synthetic isoflavone; it is methylated dehyroxydaidzein, which is metabolized to daidzein. It was developed by Chiesi Pharmaceuticals in Italy. It is marketed in the U.S., and each tablet contains 150-mg ipriflavone combined with calcium (375 mg), vitamin D (187 IU), soy isoflavones (40 mg), and 3 mg boron. The Italian product is pure ipriflavone. The recommended dose is 600 mg/day, 2 tablets b.i.d. taken with meals. Studies with ipriflavone have demonstrated prevention of bone loss over a year.^439,440,441 and ⁴⁴² Overall the effect on bone is not as great as that observed with standard doses of estrogen or bisphosphonates, perhaps not great enough to yield a benefit. A 4-year randomized trial in Europe assessed the effect of ipriflavone on bone density, urinary markers, and vertebral fractures in 474 women and could find no difference in the treated group compared with the placebo group.⁴⁴³

Phytoestrogens up-regulate cognition markers and improve memory in rats equally when compared with estrogen^444,445 There is one human study that is disturbing. Men, in a National Institutes of Health study that began in 1965, reported their tofu consumption.⁴⁴⁶ Cognition was tested in 1991-1993 when the men were age 71-93. Higher midlife tofu consumption (two or more servings per week) was associated with poor cognitive test performance, enlargement of ventricles, and low brain weight. Soy supplementation has been reported to improve measurements of memory and attention in postmenopausal women.^447,448 On the other hand, randomized trials detected no effects of soy protein, red clover, or black cohosh on tests of memory, executive function, language, visual perception, cognition, or measures of quality of life.^{391,423,449,450}

In the parts of the world where soy intake is high, there is a lower incidence of breast, endometrial, and prostate cancers.⁴⁵¹ For example, a case-control study concluded that there was a 54% reduced risk of endometrial cancer, and another case-control study indicated a reduction in the risk of breast cancer in women with a high consumption of soy and other legumes.^452,453 Daidzein and genistein urinary excretion are lower in Australian women who develop breast cancer.⁴⁵⁴ High soy and tofu consumption and high urinary excretion of isoflavones have been reported to be associated with a lower risk of breast cancer in Singapore, China, Australia, and even in American women consuming a diet rich in isoflavones.^{453,455,456,457,458} and ⁴⁵⁹ These studies have supported the belief that high phytoestrogen intake protects against breast cancer. It is by no means certain, however, that there is a direct effect of soy intake.⁴⁶⁰ Indeed, a 6-month study of the impact of administered soy protein on breast secretions in premenopausal and postmenopausal women revealed increased breast secretions with the appearance of hyperplastic epithelial cells.⁴⁶¹ Epithelial hyperplasia based on cytology in breast secretions was demonstrated in 7 of 24 (29.2%) of the subjects. Swedish and English cohort studies could not detect a relationship between dietary phytoestrogens and the risk of breast cancer.^462,463

Genistein increases epidermal growth factor in immature rat mammary tissue, and it has been hypothesized that earlier exposure to genistein promotes early cell differentiation leading to breast glands that are more resistant to the development of cancer.⁴⁶⁴ On the other hand, using the chemically induced rat breast cancer model, no evidence of isoflavone inhibition on tumor development has been detected.⁴⁶⁵ In the monkey, treated for 6 months, no proliferation was reported in either endometrium or mammary tissue.^466,467

One hypothesis speculates that phytoestrogens protect the breast by decreasing exposure to the more potent endogenous estrogens. The evidence does not support this idea. Highdose treatment (100 mg of daidzein plus 100 mg genistein) does lower estradiol and dehydroepiandrosterone sulfate levels in premenopausal women and increases cycle length.⁴⁶⁸ However, these are extremely high doses. One study reported that treatment with Asian soy foods (approximately 32 mg isoflavones per day) was associated with a 9.3% significant decrease in luteal serum estradiol levels, but there were no other changes, including follicular-phase estradiol, progesterone levels, and sex hormone-binding globulin levels, or cycle length.⁴⁶⁹ Interestingly, the reduction in luteal estradiol was observed only in Asian participants in whom urinary excretion of isoflavones was higher than nonAsians.⁴⁶⁹ These same investigators reported that a high intake of the soy protein alone (with the isoflavones removed) reduced estradiol and progesterone levels throughout the cyle.⁴⁷⁰ Other studies have found no effects on estradiol, FSH, LH, or sex hormone-binding globulin in premenopausal women,⁴⁷¹ and, most importantly, no effects on circulating hormones in postmenopausal women.^472,473 The lack of an effect on gonadotropin and steroid levels is important, depriving the clinician of a method to assess dosage.

Catecholestrogens (2-hydroxy and 4-hydroxy estrogens) have long been proposed as a metabolite pathway that could be protective, or at least antiestrogenic. Hydroxylation in the 2 or 4 position produces inactive metabolites. In 1 study, 8 premenopausal women treated with a soy milk supplement increased their urinary excretion of 2-hydroxy estrone by an average of 47%.⁴⁷⁴ Another study could detect no change in 2-hydroxyestrogens.⁴⁷¹ A study limited to Asian-American women also was unable to identify an impact of soy intake on overall estrogen metabolite excretion; however, an increase in catecholestrogens was observed with greater soy intake, balanced by a decrease in 16-hydroxylation.⁴⁷⁵

In response to soy, no significant increase in nipple aspirate levels of genistein and daidzein could be detected.⁴⁷⁶ However an indication of estrogenic stimulation occurred, as measured by pS2 (a protein up-regulated by estrogen) levels, but there was no evidence of an effect on epithelial cell proliferation, estrogen and progesterone receptors, apoptosis, or mitosis. Thus, no antiestrogenic effect could be detected, and at best there was a very weak estrogen effect. In another study, 48 women with normal breasts received a 60 g soy supplement for 14 days, and in these women lobular epithelial proliferation and progesterone receptor expression increased, an indication of estrogen stimulation.⁴⁷⁷ Some argue that the key to a beneficial impact on breast may be early exposure, and a sudden increase late in life of dietary phytoestrogens may be harmful. On the other hand, a Chinese cohort study of 5,042 breast cancer survivors documented a reduced risk of recurrence and death associated with increasing levels of soy intake, evident among women with either estrogen receptor-positive or receptor-negative disease and among either tamoxifen users or nonusers.⁴⁷⁸ In an American cohort of 1,954 breast cancer survivors, a 60% greater decrease in breast cancer recurrence was observed in postmenopausal women using tamoxifen comparing the highest level of soy intake with the lowest.⁴⁷⁹ Most of the evidence indicates that a high intake of phytoestrogens is associated with a reduced risk of breast cancer, including recurrence in breast cancer survivors. It is not known whether this effect is a marker for beneficial metabolic responses to phytoestrogens or whether there is a direct impact on breast tissue. The evidence also indicates that phytoestrogen consumption does not adversely interfere with tamoxifen’s mechanism of action.

Equol is a bacterial metabolite and the only hormonally active metabolite of the soy phytoestrogen, daidzein. It is one of the estrogenic compounds in pregnant mare’s urine, hence its name. At least in vitro, equol stimulates gene transcription with both estrogen receptors and with a greater potency than any other isoflavone.⁴⁸¹ Equol formation is totally dependent on intestinal microflora; therefore, strictly speaking it is not a phytoestrogen. To be accurate, equol is a nonsteroidal estrogen, a member of the isoflavone family, and exclusively a metabolic product of intestinal bacteria.

The most important observation regarding equol is that 30-50% of adults do not produce equol, even when challenged with high doses of soy.⁴⁸² This is in contrast to nonhuman primates and other animals; all that have been studied produce high levels of equol. Thus, there are two human populations: equol producers and equol nonproducers. The key question is whether equol producers receive greater clinical effects from phytoestrogens than nonequol producers. As noted, thus far the clinical effects of isoflavones on bone have not been impressive. In a 2-year randomized trial of postmenopausal women, isoflavone-rich soy milk increased spinal bone mass in the 45% of the subjects who were equol producers, with essentially no effect in equol nonproducers.⁴⁸³ More profound beneficial effects on the lipid profile have been reported in equol-producing women.⁴⁸² Therefore, the population destined to receive a benefit from soy intake may be limited to equol producers. Studies need to be repeated measuring the responses in individuals who are identified as equol producers or equol nonproducers. If the population destined to receive a benefit from soy intake is limited to equol producers, a convenient, inexpensive method must be developed to identify equol production. It may be possible to convert nonproducers to producers.

An emerging approach is to administer equol itself. Daidzein yields two forms of equol, the R-equol inactive isomer and S-equol, the active isomer that binds to estrogen receptor-β. S-equol has been synthesized and its administration is effective for the treatment of menopausal symptoms.⁴⁸⁴ Another alternative is the S-equol supplement made by incubating equol-producing bacteria with soy isoflavones.^485,486 S-equol also blocks the activity of dihydrotestosterone, and thus, it has potential to treat androgenic effects such as acne, hirsutism, male pattern baldness, and prostate cancer.⁴⁸⁷

Interest in estriol can be traced to Lemon’s report in 1975 that estriol limited the growth of breast tumors in the chemical-induced rat tumor model.¹⁹ However, it is usually overlooked that estradiol worked equally well in that model. Estriol treatment of postmenopausal women has no overall effect on lipids and no effect in the prevention of myocardial infarction.^488,489 Estriol, without concomitant progestin treatment, does increase the risk of endometrial cancer with the long-term oral use of 1-2 mg/day.³⁶² At least two studies have been unable to demonstrate prevention of bone loss with the administration of 2 mg estriol daily.^17,18 And one case-control study found no reduction in hip fractures with estriol compared with a lower risk with estradiol.⁴⁸⁹ There is no evidence indicating any beneficial effects unique to estriol.

Transdermal (or percutaneous) progesterone cream has been promoted to have multiple benefits. In order to achieve widespread effects, absorption must yield adequate blood levels. Two English randomized, blinded, placebo-controlled studies used 2-4 times the recommended dose and reported blood levels of about 1 ng/mL, supported by very low urinary pregnanediol levels.^490,491 An American study achieved progesterone blood levels of 2-3 ng/mL with application twice daily.⁴⁹² An Italian 1-year study did not measure blood levels but could detect no effects on bone density, lipid profiles, or depression scores.⁴⁹³

These studies indicate very little systemic absorption of progesterone from the cream product (the levels do not reach normal luteal phase concentrations), and there is great variability.

An English randomized clinical trial using transdermal doses of 5, 20, 40, and 60 mg progesterone cream could detect no differences in measures of psychological, somatic, and vasomotor symptoms compared with placebo.⁴⁹⁴ An Australian study of 16, 32, or 64 mg transdermal progesterone cream administered daily could detect no significant absorption and, most importantly, no endometrial response and no effect on flushes, lipids, bone, moods, or sexuality.^495,496 Incidentally, this study found salivary progesterone levels to be so variable that they had no meaning. Progesterone cream can produce high salivary levels, without a significant change in serum or urinary levels (the mechanism is unknown).^497,498 Red cell levels reflect serum levels and do not indicate preferential transport or sequestration.⁴⁹⁸ Clinicians and patients should be aware that transdermal progesterone cream will not reduce hot flashes more than a placebo response, but most importantly, this treatment will not protect the endometrium against the risk of endometrial cancer associated with estrogen therapy.

Wild yam creams are marketed as progesterone precursors or “balancing” formulas. Yam contains diosgenin, a plant steroid that can be converted to progesterone in a chemical laboratory but not in the human body. Predictably, a wild yam cream has no effects on a wide range of measurements in postmenopausal women.⁴⁹⁹ Some do contain progesterone, added by the manufacturer. Creams with less than 0.016% progesterone can be sold over-the-counter. There is no evidence to indicate that these preparations produce systemic effects.

Better absorption is provided by progesterone gels, an alcoholic solution with hydroxypropyl methylcellulose and water.⁵⁰⁰ With a 100 mg dose of a progesterone gel, serum progesterone levels are well into the luteal phase range, but clinical use awaits studies documenting the impact on endometrium.

Adrenal androgen production decreases dramatically with aging. The mechanism is not known, but it is not due to the loss of estrogen at menopause nor can it be reversed with estrogen treatment.⁵⁰¹ The impressive decline (75-85%) in circulating levels of DHEA that occur with aging (greater in men than in women) has stimulated a search for a beneficial impact of DHEA supplementation.⁵⁰²

The only proven function of DHEA and its sulfate, DHEAS, is to provide a pool of prohormone for conversion to androgens and ultimately estrogens. By age 70 or 80, the circulating levels in men and women are about 10% of peak levels that occur between 20-30 years of age. DHEA supplementation does not produce improvements in menopausal symptoms, mood, libido, cognition, or memory, but it does increase testosterone and decrease HDLcholesterol.⁵⁰³ The acute administration of DHEA did produce a modest effect on sexual response in postmenopausal women, but the dose was enormous, 300 mg.⁵⁰⁴

Although low levels of DHEA and DHEAS have been reported to be associated with increased risk of cardiovascular disease in men, in women conflicting results are found in cross-sectional data. In a longitudinal study of 236 women, higher levels of DHEA and DHEAS in middle-aged women correlated with an increased risk of cardiovascular disease.⁵⁰⁵

DHEA supplementation, 50 mg/day, produced reproductive levels of DHEAS in elderly men, did not change levels of testosterone and dihydrotestosterone, and raised estradiol and estrone levels, although still within normal range.⁵⁰⁶ In women, 25 or 50 mg/day increased testosterone levels, decreased sex hormone-binding globulin levels, and produced adverse effects on the lipid profile.^507,508 Exogenously administered DHEA is converted to potent androgens and estrogens. Potential long-term effects include hirsutism, alopecia, voice changes, prostate and breast effects, and an increased risk of coronary heart disease. Supplementation with DHEA requires titering of dosage using the circulating level of total testosterone and keeping the concentration below 80 ng/dL. This is difficult because the U.S. Food and Drug Administration measured the DHEA content of 45 commercial products, and assayed values varied from 0 to 109.5%!⁵⁰⁹

The daily intravaginal use of DHEA in low doses is reported to improve vaginal atrophy and sexuality, with little change in the serum levels of DHEA, estradiol, and testosterone.^510,511,512 and ⁵¹³ Presumably, the DHEA is converted locally to estrogen and testosterone. Studies documenting the effects on target tissues, such as endometrium, bone, and liver, will be required to assess the long-term safety of this treatment.

The thickness of the postmenopausal endometrium as measured by transvaginal ultrasonography in postmenopausal women correlates with the presence or absence of pathology. However, the severity of pathologic change does not correlate with the measured thickness.⁵²⁶ Endometrial thickness (the two layers of the anterior and posterior walls in the longitudinal axis) under 5 mm is reassuring and allows conservative management.^{527, 528} Endometrial thickness greater than 4 mm requires biopsy; it is estimated that 50-75% of bleeding patients on hormone therapy and evaluated by ultrasonography will require biopsy.^526,529 An endometrial thickness less than 5 mm in women receiving hormone therapy, either a sequential regimen or a daily combination of estrogen-progestin, is reassuring.^528,530,531 It seems logical that endometrial thickness by ultrasonography in patients on a sequential regimen can be affected by day in the treatment cycle, and for that reason, ultrasonography assessment should be obtained toward the end of the progestin phase or at the beginning of the cycle.^532,533 and ⁵³⁴ An Italian study concluded that endometrial thickness measured soon after withdrawal bleeding in women on a sequential regimen was comparable to thickness in women on a continuous, combined program of estrogen-progestin treatment.⁵³⁵ When a thick endometrium is associated with atrophic endometrium on biopsy, polyps are often present. Greater accuracy can be gained by the instillation of saline into the uterine cavity during ultrasonography.⁵³⁶ Doppler velocimetry does not improve the accuracy of discriminating between normal and abnormal endometrium.⁵³⁷ A clinician should not be satisfied with “normal” findings on ultrasonography if a patient has persistent bleeding. The pursuit of abnormal bleeding despite “normal”
findings should reduce missed cases of pathology to nearly zero.⁵³⁸ In this circumstance, hysteroscopy is recommended.

The administration of a progestational agent (e.g., 10 mg medroxyprogesterone acetate for 2 weeks) was developed by R. Don Gambrell Jr. as a means of detecting the presence of estrogen-dependent endometrium in postmenopausal women.⁵³⁹ A withdrawal bleed would indicate that an endometrial response has occurred to the progestin, a response that requires previous endometrial stimulation by estrogen and indicates the need for endometrial assessment. In other words, the lack of a withdrawal bleed is reassuring for clinician and patient. Concern with this clinical maneuver has focused on whether there are falsenegative and false-positive responses. Several studies are now available regarding the efficacy and validity of this method.^540,541 and ⁵⁴² The published data indicate that most, but perhaps not all, women with endometrial proliferation, hyperplasia, and even cancer will respond with a withdrawal bleed after a progestin challenge, and ultrasonography measurement of endometrial thickness will be greater than 4 mm.^543,544 In an experiment in monkeys, 1 of 14 animals treated with estrogen did not bleed and 5 of 13 placebo-treated animals did bleed.⁵⁴⁵ The problem is that the studies thus far do not consist of very large numbers, and there is a lingering question whether a patient with abnormal endometrium will always bleed in response to progestin treatment and withdrawal.