The Hormone Continuum: Accrual of Women's Health Benefits

Three important issues in women's health care are oral contraceptives (OCs), the management of perimenopausal changes in endocrine physio-logy, and postmenopausal hormone replacement therapy (HRT). All involve the use of both estrogen and progestin. Approximately 80 percent of American women take OCs at some time in their lives, and approximately 30 percent at least start HRT. Decisions about these three issues entail a careful risk/benefit analysis involving the individual patient's risk factors and a careful weighing of benefits and long-term risks to breast, uterine, ovarian, cardiovascular, and cerebrovascular health. This, in turn, requires evidence-based medical practice derived from epidemiological studies and clinical trials, among which there are often disparate findings.

This program was jointly sponsored by the University of Minnesota Office of Continuing Medical Education and Health Learning Systems, and was funded by an unrestricted educational grant from Ortho-McNeil Pharmaceutical, Inc.

Risk/Benefit Considerations for Hormonal Therapy: OCs and HRT

Using the categorization schemes of evidence level and making evidence-based clinical recommendations shown in Tables 1 and 2, Daniel R. Mishell, Jr, MD (University of Southern California) evaluated current safety evidence on OCs and HRT.

OCs
Although there is evidence that estrogen and progestin both cause peripheral insulin resistance, it does not appear to occur with the lower doses of steroids and hormones used in contemporary OC formulations (less than 50mcg of estrogen). Additionally, they can have a beneficial effect on the lipid profile by raising HDL, lowering LDL, and mildly increasing triglycerides. OCs are contraindicated, however, for women who have diabetes with vascular disease.

Cigarette smoking potentiates the effect of OCs on the coronary arteries and increases the risk of myocardial infarction. OCs should not be prescribed for women over 35 years of age who smoke. Uncontrolled hypertension is also a contraindication.

Studies in which all data were collected after 1980 have consistently found that healthy women who use OCs have a statistically significant three- or four-fold increased relative risk of venous thrombosis. As a result, the strength recommendation as to the relative risk of venous thrombosis from the use of OCs is A. Despite the elevated relative risk, however, the absolute risk is low, at about one in 10,000 women per year. Hence the World Health Organization (WHO) does not recommend routine screening for thrombophilic disorders prior to OC use (strength recommendation B). Four studies have concluded that the risk of venous thrombosis from OC use is not dependent upon estrogen dose below 50 mcg, so there is a strength recommendation of A that there is no difference in risk between 20 and 50mcg with the same dose progestin.
WHO finds that current or prior OC use is not associated with increased risk of myocardial infarction in nonsmokers (evidence level II-2). Smoking is the most important independent risk factor that synergizes with OC use. Nonsmoking, normotensive, nondiabetic women of any age who use OCs do not have an increased risk of myocardial infarction. OCs containing 35 mcg of ethinyl estradiol or less may be used safely by nonsmoking women of any age. Smokers over 35 years of age should use nonestrogen or nonhormonal contraceptive methods. The strength of these recommendations is A.

The incidence of stroke is increased with first-generation, high-dose OCs. Three studies have found, however, that OCs containing less than 50 mcg of ethinyl estradiol have virtually eliminated the increased risk of both ischemic and hemorrhagic stroke in nonsmoking, normotensive women who do not have other risk factors. There is no evidence, however, that lowering the dose to 20 mcg reduces risk further. The evidence level of these findings is II-2.

A large collaborative study has shown that current users of OCs had a 1.24 relative risk of diagnosis of breast cancer. However, ten or more years after discontinuing the use of OCs, former users have no higher risk of breast cancer than women who never used them. This suggests that estrogen probably does not cause breast cancer. One plausible explanation of this apparent contradiction is that estrogen may promote the growth of small occult tumors and increase the chance of diagnosis without having been oncogenic. There is no evidence that a family history of breast cancer should contraindicate OC use.

Several studies indicate that the use of OCs decreases risk of both endometrial and ovarian cancer, that the protection persists after discontinuation, and that the duration of protection against ovarian cancer correlates directly to the duration of OC use. Users of OCs incur no increased risk of cervical dysplasia or squamous cell carcinoma, but there is an increased risk of adenocarcinoma of the cervix.

HRT
Table 3 summarizes the evidence levels and strengths of recommendations regarding the health benefits (decreased relative risk) from HRT. Dr. Mishell emphasized that HRT reduces the risk of hip fracture by approximately half, and that a meta-analysis of 31 studies of the effects of estrogen replacement therapy (ERT) alone consistently found a risk reduction for cardiovascular disease, principally myocardial infarction, of approximately half. With unopposed estrogen, however, risk for endometrial cancer increases two- to eight-fold (strength of recommendation A; evidence level II-1). HRT with sequential or cyclic progestin and estrogen alone, however, reduces the relative risk for endometrial cancer to approximately 2.0; and continuous combined progestin/estrogen actually reduces the overall risk. Unopposed estrogen also appears to increase risk for venous thrombosis and pulmonary embolism (strength of recommendation B; evidence level II-2).

More than 50 studies and six meta-analyses have found no relationship between HRT and breast cancer, though a significantly increased risk for breast cancer diagnosis - 25 to 30 percent - has been observed in long-term users of estrogen only. The Breast Cancer Diagnosis and Detection Project (77 cases) found that a 40 percent increase in risk for breast cancer diagnosis was limited to long-term current users with body mass index of less than 24.4. A Los Angeles County study found no increased risk with long-term current use or past use up to 15 years. Women on HRT for more than 15 years had an increased risk of carcinoma in situ, but not of invasive cancer. A study conducted in Washington found no increase in invasive ductal cancer with either estrogen alone or combined with a progestin in either current or prior users. Dr. Mishell concluded that current or prior use of HRT incurs no added risk of breast cancer (recommendation A) except for current long-term use of estrogen and cyclic progestin (recommendation B). Several studies have confirmed a significantly reduced risk of breast cancer mortality among women diagnosed while taking HRT.

Table 1. Categorizing the Level of Evidence

I.          At least 1 properly controlled randomized trial

II–1.     Controlled trials without randomization

II–2.     Well-designed cohort and case-control studies

II–3.     Cross-sectional studies, studies with external control groups, or ecological studies

III.        Evidence derived from report of an expert committee, which itself used a scientific approach
CMAJ. 1979;121:1193.

Table 2. Making Evidence-Based Recommendations

A. There is good evidence for cause and effect
B. There is fair evidence for cause and effect
C. There is insufficient evidence
    - C1. No evidence or data
    - C2. Evidence conflicting and unresolvable
    - C3. Two or more plausible explanations
D. There is fair evidence against cause and effect
E. There is good evidence against cause and effect

Table 3. Evidence-Based Benefits of ERT

Level Strength of Decreased RR of Evidence Recommendation

Vasomotor symptoms                   I A
Genitourinary symptoms             II-2 A
Cardiovascular disease               II-2 A
Hip and vertebral fractures         II-2 A
Alzheimer's disease                     II-2 B
Colorectal cancer                         II-2 B
Tooth loss                                     II-2 B
Breast cancer mortality               II-2 B
Death age < 80                             II-2 B

Current Perspectives on Oral Contraceptives

Half of all unintended pregnancies occur in women who use contraception. Many occur in women over the age of 40 years who never used OCs or who discontinued them prematurely, often because of side effects such as breakthrough bleeding, spotting, nausea, heavy periods, or amenorrhea. Older women also tend to be more concerned about long-term health risks from taking OCs, though their fears may not be based on fact. Nevertheless, while 21 percent of American women between 35 and 39 years of age use OCs, use drops sharply thereafter. The rates of abortion and complicated pregnancy are high in this population. Fifty-one percent of all pregnancies in women over age 40 are unintended, and 65 percent of them are voluntarily aborted. In addition to ensuring effective contraception in the final reproductive years, OCs impart a number of important health benefits of which many women are unaware. Jo Ann Rosenfeld, MD (Johns Hopkins University) emphasized a few of them.

A study of 201 women who took triphasic norgestimate with 35 mcg of ethinyl estradiol for dysfunctional uterine bleeding - abnormal bleeding from the uterus not attributable to pelvic pathology or systemic disease - demonstrated significant improvement compared to placebo in symptoms and improvement in quality of life after three cycles. In a case-controlled analysis of hospital data, women taking OCs had a 60 percent reduction in hospitalizations for pelvic inflammatory disease, irrespective of dose. All currently available mono-phasic OCs significantly reduce risk for functional ovarian cysts, the fourth leading cause of gynecologic hospitalizations. The incidence of ectopic pregnancy has been shown to decline from 2.6 per 1,000 women annually who use no contraception to 0.005 per 1,000 OC users per year. The use of OCs reduces the risk of ovarian cancer between 40 percent and 80 percent, and the risk for endometrial cancer by 50 percent. Risk reduction increases the longer the woman uses OCs.

Benefits reach beyond the pelvic organs. For example, a case-controlled study demonstrated that the risk for benign breast disease is reduced for the entire duration of OC use in women who start OCs prior to the first full-term pregnancy. Long-term use of OCs during reproductive years may increase peak bone mass density, allow women to enter menopause with healthy bones, and reduce the prospect for subsequent osteoporotic fractures. Use of OCs in a trial of 4,000 women was associated with a decreased risk of developing colon and rectal cancer. A meta-analysis of nine hospital- and population-based studies suggests that OC use may prevent progression of rheumatoid arthritis; and a case-controlled study found that OC users have a relative risk of developing rheumatoid arthritis of 0.1, suggesting a protective effect.

Notwithstanding all of these apparent advantages of oral contraception, even at current low doses (no dose higher than 50 mcg is available in the United States), some women experience side effects. More controversial, though, is the question of minimizing side effects by reducing dose while preserving efficacy. Although most current formulations are in the 25 to 35 mcg range, 20 mcg pills have appeared. The low ethinyl estradiol dose has been shown to be associated with increased follicular activity, however, raising the possibility that it might not always suppress ovulation. Moreover, in one study, cycle control was superior in women who took 35 mcg doses than in women taking 20 mcg. In a six-cycle multicenter open-label, randomized trial, there was significantly less breakthrough bleeding with the 35 mcg dose than with 20 mcg.

Transition from OCs to HRT

Ronald T. Burkman, MD (Tufts University) defined the perimenopausal period, a time of extreme changes in hormone physiology, as the five to seven years preceding the final menstrual period and one to two years following it. During these years, many of the follicles that remain in the ovaries become atretic, many cycles are anovulatory, and a reduction in progestin output during the luteal phase causes wide variations in cycle length. Duration and quality of bleeding may also be affected. OCs are effective for managing the benign endometrial disorders that occur during these years, and they should probably be considered first-line therapy for these anovulatory types of bleeding in order to stabilize the endometrium. They also provide secondary benefits appropriate to the health concerns of mature women as outlined by Dr. Rosenfeld. Moreover, they decrease both the number and severity of hot flashes in perimeno-pausal women. Two studies have shown that contrary to popular notion, OCs do not induce weight gain.

Converting women to HRT at age 47 or 48 years who have irregular cycles and hot flashes is not advisable. Although the lower-dose estrogen in HRT than in OCs potentially reduces the risk of venous thrombosis, it does not suppress ovulation, so HRT provides no contraception. Moreover, it does not prevent unpredictable dysfunctional uterine bleeding, and may aggravate it.

Monitoring follicle stimulating hormone (FSH) levels during the hormone-free interval of OC use is not a reliable way to measure progression toward menopause. Seamless transition to HRT from OCs as women approach their mid-fifties, after five to ten years of treating for symptoms with 25 to 35 mcg doses, is safer and more beneficial.

Current Perspectives on HRT Regimen

Only 20 to 25 percent of postmenopausal American women are on HRT. The fact that the majority of these have had hysterectomies suggests that side effects, bleeding, and fear of hormone-induced uterine cancer are major deterrents among women who retain their uteri. As reported by John H. Mattox, MD (University of Arizona), 34 percent of postmenopausal women list health and fitness as their primary concern, yet many are either uninformed about the benefits of HRT or are deterred from using it by misplaced fears of long-term risks. In fact, the list of long-term benefits continues to grow. Recently it has been shown, for example, that HRT may reduce senile macular degeneration and that it may improve immune responsiveness by restoring menopause-depressed levels of natural killer cell activity. Evidence from studies in which users and non-users of HRT were followed for as long as 23 years indicates overall that HRT prolongs lives by a mean of approximately three years.

Increasingly, women are requesting continuous combined therapy rather than cyclic or sequential dosing of estrogen and progestin. It is well established that estrogen alone produces the most advantageous lipid profile, a surrogate marker for cardiovascular protection. But estrogen unopposed by progestin elevates the risk for endometrial carcinoma. Since many of the side effects from HRT are attributable to the progestin component, the pharmaceuticals industry is searching for new ways to make it more tolerable. Norgestimate (NGM) has an excellent progestogenic profile, very strong affinity for the progesterone receptor, minimal affinity for androgen receptors, and a favorable metabolic profile. As for estrogen, the strongest case can be made for 17b-estradiol (17b-E2). It occurs naturally in humans, but can be derived from plants for use in HRT preparations. It has been used and studied extensively, and it has therapeutic efficacy comparable to other estrogens.

Recently, based on evidence regarding receptor dynamics, the concept of constant estrogen and intermittent progestin has been introduced. Progestin is taken for three days, discontinued for three days, and then resumed for three days while taking estrogen continuously. When estrogen is taken alone, the estrogen and progestin receptors are up-regulated, thereby potentiating maximum estrogen benefits and increasing sensitivity to progestin. In the following three days of co-therapy, both receptors are down-regulated, minimizing progestin over-exposure. When this is done with longer-half-life progestins, the total progestin load per month is reduced and with that, potentially, side effects as well. This should encourage improved compliance. Compared with one of the commonly prescribed triphasic oral contraceptives, this intermittent method achieves protective endometrial effect with approximately a quarter of the progestin dose. A recent clinical trial comparing an intermittent 17b-E2/NGM preparation (Ortho-PREFEST®) with other preparations and schedules has demonstrated significant reductions in side effects, preservation of metabolic benefits, and virtual elimination of endometrial neoplasia. (See Table 4.) Less than one percent of trial subjects withdrew because of bleeding associated with this preparation.

Table 4. Continuous 17b-E2/Intermittent NGM: Endometrial Biopsy Results After 1 Year

	17b-E2	17b-E2/NGM

Total subjects	265	242

Total evaluable biopsies	256 (97%)	227 (94%)

Normal endometrium	182 (71%)	227 (100%)

Simple hyperplasia	64 (25%)	0 (0%)

Complex hyperplasia	2 (0.8%)	0 (0%)

Hyperplasia with cytological atypia	8 (3%)	0 (0%)

Ortho-PREFEST® prescribing information. Raritan, NJ:
Ortho-McNeil Pharmaceutical; December, 1999.

On October 11, 2000, Grabrick et al published results of a study involving breast cancer risk among women with familial predisposition who took OCs prior to 1975, when standard doses were 50 mcg of EE or higher. (JAMA. 2000;284:1791.) Data indicated a significant risk increase in this population. No increased risk was found, however, in women whose first-degree relatives had breast cancer but who used OCs exclusively after 1975. The study confirmed prior evidence that current OCs are associated with little or no risk of breast cancer among average-risk women. OC use is known to reduce risk of ovarian cancer, including in women with BRCA1/2 mutations. In high-risk women, decisions about OC use need to be made on an individualized basis with consideration to baseline risk, alternate cancer risk reduction strategies, and patient preferences.

Return to Primary Care Symposium