A 60-Year-Old Woman Trying to Discontinue Hormone Replacement Therapy

DR REYNOLDS: Mrs W is a 60-year-old white woman who is trying to discontinue her estrogen therapy. She lives alone near Boston and teaches middle school. Her 29-year-old son lives in another state. She has managed care insurance.

At the age of 40, Mrs W underwent hysterectomy and bilateral salpingo-oophorectomy to treat refractory endometriosis. After several days of severe flushes, sweats, and depression, she began taking conjugated estrogen, 0.625 mg/d. With that therapy she has felt well for the past 2 decades. Over the past few years, she became increasingly concerned about her family history of breast cancer and the health risks of postmenopausal hormone replacement therapy (HRT) after reading about it in the lay press. Mrs W asked her former physician about these risks, and he encouraged her to continue treatment. At times, she discontinued the estrogen on her own, but disabling hot flushes ensued and she restarted treatment each time.

Four months ago, Mrs W visited a new primary care physician, Dr K, who encouraged her to taper the HRT. With her dose decreased to 0.3 mg/d, she didn't feel as well as she did with the higher dose. She then stopped taking the medication according to the planned taper. She has numerous hot flushes daily, sometimes as often as every 30 minutes. The hot flushes wake Mrs W from sleep. She is not depressed, but she has a general sense of not feeling as well as when taking the hormonal therapy.

Her past medical history is significant for multiple breast cysts requiring aspiration and an excisional biopsy of a left breast mass many years ago that was benign. She had atrial fibrillation treated with nodal ablation and is now in sinus rhythm. She had superficial phlebitis 1 year ago. She has never had a venous thromboembolic event. Her menarche was at age 10; her only pregnancy was at the age of 32 and was carried to term.

Mrs W's mother had rectal cancer diagnosed at the age of 70. Her maternal grandmother was in her 50s when she died of breast cancer; Mrs W's younger sister was diagnosed as having breast cancer in her late 30s. Her father died of an abdominal aortic aneurysm. There is no family history of osteoporosis or premature coronary heart disease (CHD).

Mrs W does not smoke or drink alcohol. She walks several times a week and has begun strengthening exercises. Her husband died several months ago and she reports "normal grief" and a strong support network.

On physical examination, Mrs W looked younger than her age. At 5'3" she weighed 116 lb; her blood pressure was 120/60 mm Hg and her pulse was regular at 60/min. She had a left breast scar at 3-o'clock but otherwise normal breast examination findings. Her pelvic examination showed no cervical cuff and no adnexal masses. The remainder of her physical examination results were normal. Mrs W's only medication was atenolol, 25 mg/d.

Laboratory data showed a total cholesterol level of 179 mg/dL (4.6 mmol/L) and high-density lipoprotein (HDL) level of 91 mg/dL (2.4 mmol/L), measured in August 2000. Bone density testing in June 2001 showed a T score at the femoral neck of −1.68. A mammogram within the last year was unremarkable.

I have been on [hormone therapy] for about 20 years. But the more I read, it's about time I have to get off this. I tried stopping it, but I am so uncomfortable. Without asking anybody, I just stopped taking the hormone therapy and then of course I would become depressed, and then I'd put myself back on it.

I continued to take it only because I couldn't stand being off the hormone. I really couldn't function. When I was on the 0.3 [mg/day], I felt pretty good—not as good as when I was on the 0.625 [mg/day]—but not as badly as I feel now.

Ideally, when I sit down by myself and think about it, I would love to be off them. I would love to get rid of these hot flushes. [A hot flush is when] you feel very, very warm and it starts becoming very prickly, works its way up from your toes through your entire body and then manifests itself in your face . . . and I'm dripping with perspiration. It probably lasts for about 5 minutes. The last week or so, it's been happening about every half hour.

It just seems like each day I am getting more and more uncomfortable. Right now I'm ready to just scream and say, oh please just put me back on it.

At this juncture, someone like me who has been taking hormones for so long but still has symptoms of menopause, what can I do? There must be something [women] can take that would alleviate some of the symptoms. That's what I would like more than anything, so I could just go on and be relatively happy—and not so warm!

My concerns with Mrs W were, I think, similar to hers, that she'd been on hormone therapy for 20 years. She has risk factors for breast cancer and a family history, and I agreed with her that there was a concern about continuing it.

We made the decision to go ahead and try to taper her off the HRT. While coming off of it, although she did it much more slowly this time, she again became quite symptomatic and found it very difficult to tolerate the symptoms.

I would like to know if Dr Grady thinks that we should proceed with trying to discontinue Mrs W's hormone therapy. How would she suggest handling her low bone density, and in the setting of wanting to discontinue the hormone therapy, what alternative regimens would she suggest? Finally, does Dr Grady have any other suggestions for how to manage the patient's symptoms once HRT is discontinued?

Was it appropriate to treat Mrs W with hormone therapy? Should she continue taking hormones? Which women should discontinue postmenopausal hormone therapy? Are there alternatives to hormone therapy? Can we predict which women will have trouble stopping hormone therapy? What is the best way to stop hormone therapy to minimize symptoms? What do you recommend for Mrs W?

DR GRADY: Mrs W has been taking postmenopausal hormone therapy for many years. She started taking it for classic menopausal symptoms after hysterectomy and oophorectomy at age 40 and did well with this treatment for 20 years. After so many years, is there an indication for Mrs W to continue taking estrogen?

Estrogen use is associated with common adverse effects such as bleeding and breast tenderness and more serious ones including increased risk for venous thromboembolism,¹ gallbladder disease,² and probably breast cancer.³ The indications for using a drug with known adverse effects should be clear. There are 2 general indications for HRT: treatment of menopausal symptoms⁴ and prevention of osteoporosis.^{5 - 6}

Hot flushes, sweats, and insomnia are common around the time of menopause.⁷ In women like Mrs W who undergo bilateral oophorectomy, vasomotor symptoms are often more severe than in women who undergo natural menopause,⁸ perhaps because of the sudden and precipitous drop in estrogen associated with removal of the ovaries. Estrogen is the most effective therapy available for treatment of vasomotor symptoms (Table 1)^{9 - 12} and was clearly indicated for Mrs W after her surgery. For treatment of symptoms, all types and routes of administration of estrogen are equally effective.⁴ Some women, especially younger women who have undergone oophorectomy, may require a higher than average dose of estrogen to suppress symptoms. In contrast, many women with flushes at the time of menopause do well with a lower dose of estrogen, such as 0.025 mg/d of transdermal 17β-estradiol¹³ or 0.3 to 0.4 mg/d of oral conjugated estrogen.¹⁴ The new selective estrogen receptor modulators (SERMs) such as tamoxifen and raloxifene can cause or exacerbate hot flushes in 10% to 30% of women.^{15 - 16} In contrast to the marked beneficial effect of HRT on flushes in younger women, there is little evidence that HRT improves quality of life in older, asymptomatic women.¹⁷

Mrs W reported that HRT improved the depressive symptoms she experienced after oophorectomy. Postmenopausal hormone therapy should not be used for treatment of major depression¹⁸ but might improve depressive symptoms in menopausal women, perhaps by relieving flushing and improving sleep.¹⁹ Hormone therapy has not been shown to be effective for other symptoms of menopause, such as labile mood, memory loss, and decreased libido.²⁰

In most women, menopausal symptoms are transient. About 30% to 50% of women improve within several months,^{4 ,14} and flushes resolve in most women within 4 to 5 years.^{7 ,21} However, the natural history of flushes is highly variable and some women continue to have flushes for many years after menopause. Most women who begin estrogen for treatment of flushes can taper or discontinue a few years after starting, but some, like Mrs W, will continue to be symptomatic for many years. It is my practice to attempt to stop hormone therapy about 2 or 3 years after starting unless there is an indication to continue for prevention of osteoporosis, such as prior postmenopausal fracture or bone density T score lower than −2.5.²²

Unopposed estrogen markedly increases the risk of endometrial cancer unless accompanied by a progestin.²³ Hormone therapy also increases the risk of venous thromboembolism about 3-fold¹ and gallbladder disease about 40%.² Other than these conditions, adverse effects should not be a major concern when using short-term hormone therapy around the time of menopause. If treatment can be tapered and stopped within about 5 years, there is no evidence of increased risk of breast cancer.^{3 ,24}

Should Mrs W continue hormone therapy for prevention of disease? For disease prevention, long-term or even lifelong treatment is probably necessary,²⁵ and potential adverse effects increase with long-term use. Hormone use for longer than 5 years is associated with a 30% to 60% increased risk of breast cancer.^{3 ,24} In addition, the absolute risk of breast cancer and venous thromboembolism increases with age, so that any increased relative risk attributable to HRT results in a much higher risk of these adverse outcomes in elderly women. Thus, when treating women like Mrs W, the benefits of hormone therapy for prevention should outweigh the risks. In my view, this requires evidence from large randomized trials.

Prevention of CHD. Although Mrs W's medical history and family history do not put her at high risk for CHD, a healthy woman her age has about a 31% lifetime risk of dying of heart disease.²⁶ If HRT prevented CHD, it would have a large public health benefit. Multiple observational studies have found that women who use HRT have a 35% to 50% lower risk of coronary disease than nonusers.^{26 - 27} These findings are supported by plausible biological mechanisms, including beneficial changes in low-density lipoprotein and HDL cholesterol,²⁸ improved endothelial function,²⁹ and slower progression of atherosclerosis in animals.³⁰ However, the results of recent clinical trials conflict sharply with these findings (Table 2). The Heart and Estrogen/progestin Replacement Study (HERS), a randomized, blinded, placebo-controlled trial among 2763 postmenopausal women with documented CHD, found no overall benefit of 4 years of treatment with estrogen plus progestin, and women randomized to HRT had a 50% increase in risk of CHD events during the first year of therapy.³¹ It has been suggested that the negative findings of this trial were due to inadequate power to detect a benefit, use of a progestin with estrogen rather than estrogen alone, and inclusion of women with preexisting CHD rather than healthy women. In fact, confidence intervals around the main finding (relative risk, 1.0; 95% confidence interval, 0.80-1.20) demonstrate that it is highly unlikely that a true reduction in CHD risk of 20% or greater was missed.³¹

Since publication of HERS, other major trials have provided evidence concerning effects of unopposed estrogen and of HRT in women without preexisting CHD (Table 2). The Estrogen Replacement in Atherosclerosis trial found that neither unopposed estrogen nor estrogen plus progestin slowed the progression of atherosclerosis measured by quantitative angiography among 309 women with preexisting CHD.³² The Women's Health Initiative, a randomized trial among more than 27 000 women, 98% of whom did not have preexisting CHD at the start of the trial, recently reported that both unopposed estrogen and estrogen plus progestin were associated with an increased risk of cardiovascular events compared with placebo during the first 3 years of treatment.^{33 - 34} The Women's Estrogen for Stroke Trial also found no reduction in risk of stroke and death or coronary events among 664 postmenopausal women treated with 17β-estradiol or placebo for 2.8 years. Women in the hormone group had a 2.3-fold increased risk of stroke in the first 6 months of the treatment.³⁵ There is currently no convincing evidence that tamoxifen^{16 ,36} or raloxifene³⁷ cause an early increased risk of CHD or any overall reduction in risk. A clinical trial among more than 10 000 women is in progress to assess the effect of raloxifene on CHD risk.³⁸ Although HRT^{26 ,35 ,39} and raloxifene³⁷ do not appear to alter the risk of stroke, tamoxifen was associated with a nonsignificant 60% increased risk of stroke in the Breast Cancer Prevention Trial.¹⁶

What explains the difference in the findings of observational studies and the randomized trials described above? Even a very well-conducted observational study may produce the wrong answer if there are unmeasured differences between hormone users and nonusers. Women who take HRT are generally healthier and wealthier than nonusers.^{40 - 41} To be classified as a hormone user in an observational study, a woman must not only be prescribed estrogen, but get the prescription filled and continue to take the medication. Adherence—taking a medication as instructed—has been shown to be a strong marker for low risk of coronary events, even when adherence is to a placebo.^{42 - 43} Unmeasured differences between hormone users and nonusers due to selection or adherence may account for the lower risk of CHD among hormone users in observational studies. Randomized assignment to hormone use in a clinical trial eliminates these and other potential baseline differences between the treatment groups and is the reason that the evidence from randomized trials is more reliable than that from observational studies.

In short, no randomized trial of hormone therapy has shown a reduced risk of coronary events in any group of postmenopausal women. This lack of clinical trial evidence recently led the American Heart Association to revise guidelines to recommend against initiating use of HRT for secondary prevention of cardiovascular disease.³⁸

Prevention of Osteoporotic Fractures. For Mrs W and many postmenopausal women, maintaining healthy bone density is an important consideration. Hormone therapy reduces bone loss in postmenopausal women. After 2 to 3 years of treatment, clinical trials show 5% to 7% greater spinal bone density in women randomized to treatment with HRT compared with those randomized to placebo.⁵ Observational studies suggest an approximately 50% lower risk of hip and other fractures among women who use HRT compared with nonusers.²⁶ Based on these data, estrogen has long been first-line therapy for prevention of osteoporotic fractures. However, no large randomized trial has been completed to support the observational findings regarding fracture risk. In a recent meta-analysis of 22 small trials, all but one of which was conducted to examine nonfracture outcomes, women randomized to hormone therapy had a 27% lower risk of osteoporotic fracture compared with those randomized to placebo.⁶ In HERS, the largest trial to date with fracture outcomes, women who took hormones for a mean of 4.1 years did not have a lower risk of any type of fracture compared with women who took placebo.⁴⁴ Estrogen therapy may not have reduced the risk of fracture among women in HERS because they were not, on average, osteoporotic, and drugs appear to be much more effective in reducing fracture risk in women with osteoporosis than in those with normal bone density.⁴⁵ Estrogen is approved by the US Food and Drug Administration (FDA) for the prevention of osteoporosis, but due to the weakness of the evidence, not for the treatment of osteoporosis.

Bisphosphonates including alendronate and risedronate have been shown in randomized trials to preserve bone density about as well as estrogen^{46 - 47} and to reduce the risk of hip and other fractures by 40% to 50% in osteoporotic women.^{46 - 48} These drugs are associated with gastroesophageal symptoms⁴⁹ and occasionally cause esophagitis, but they are otherwise free of adverse effects. Raloxifene also has been shown in a large trial to decrease the risk of spinal fracture by about 40%,⁵⁰ but the risk of hip and other clinical fractures was not reduced and raloxifene should not be considered a first-line drug for prevention of fractures in women with osteoporosis.

Several lines of evidence suggest that estrogen use increases the risk of breast cancer, about which Mrs W is concerned. Estrogen promotes tumor induction and growth in animals^{51 - 52} and increases the proliferation rate of human breast cancer cells in vitro.⁵³ Many of the traditional breast cancer risk factors, such as early menarche, late menopause, and nulliparity, are related to prolonged exposure to endogenous estrogen. Higher endogenous estrogen levels⁵⁴ and higher bone density (a marker of long-term endogenous estrogen exposure) are associated with increased risk for breast cancer.⁵⁵ Finally, multiple observational studies have found an increased risk of breast cancer among long-term hormone users. Pooled analyses of 51 published observational studies show no increase in risk of breast cancer among women who took estrogen for less than 5 years.³ This would suggest that women who take estrogen for a few years for treatment of symptoms do not have an increased risk of breast cancer. With longer duration of use, however, the risk was observed to increase 30% to 60%.³ Using population-based estimates of breast cancer,³⁶ an additional 2 to 4 cases would be expected per 1000 women aged 65 to 69 years treated with estrogen for 10 to 15 years. No randomized trial has addressed the risk of breast cancer among estrogen users, so it is possible that the observed increased risk is due to other differences between estrogen users and nonusers. However, because physicians tend to avoid prescribing estrogen to women at high risk for breast cancer, the actual relative risk for breast cancer among hormone users could be higher than observed.

Several years ago, I thought that preventive hormone therapy should be prescribed to most postmenopausal women except those at high risk of breast cancer.⁵⁶ This approach was based on the belief, supported mainly by observational evidence, that estrogen use reduced risk for both osteoporotic fractures and CHD. Because fractures and CHD are common and life-threatening, the net effect of reducing risk for both diseases would be beneficial, even if associated with an increased risk for venous thromboembolism and breast cancer.²⁶ Recent evidence that HRT does not reduce risk of CHD and the availability of other effective treatments for prevention of osteoporotic fractures have drastically altered my approach, as well as that of others.⁵⁷ Rather than prescribe preventive hormone therapy for most postmenopausal women, I currently recommend it only for women at high risk of fracture who are not at high risk for breast cancer. The evidence that bisphosphonates reduce risk of hip and other fractures is stronger than the evidence supporting use of estrogen for this purpose, and these drugs also have fewer potential serious adverse effects. For these reasons, bisphosphonates should be the first choice of therapy for prevention of fractures, especially in women at high risk of breast cancer.

What about preventive hormone therapy for Mrs W? She has no particular risk factors for osteoporotic fracture except that she is white. Her bone density T score was −1.68 at the lumbar spine, which is osteopenic but not osteoporotic. Following the recommendations of the National Osteoporosis Foundation,²² I would recommend that she take 1200 mg/d of supplemental calcium plus 400 to 800 IU/d of vitamin D but no other drug therapy for prevention of fractures unless her bone density falls below a T score of −2.0. In contrast, she is at high risk for breast cancer given her age, strong family history, prior breast biopsies, early menarche, and late pregnancy. Based on a risk model developed by the National Cancer Institute,⁵⁸ her risk of breast cancer is about 4.8% over the next 5 years, compared with a 1.8% risk in a woman her age without additional breast cancer risk factors (tool available at http://bcra.nci.nih.gov/brc/q1.htm). Given that Mrs W is not at particular risk for fracture but is at high risk for breast cancer, I would agree with her current primary care physician that she should stop her hormone therapy if possible. I generally try to avoid prescribing hormone therapy for women who have a first-degree relative with breast cancer and those with a greater than 2% risk of developing breast cancer in the next 5 years.

Unfortunately, there is very little evidence to guide Mrs W and her physician regarding how to stop HRT. We do not know what proportion of women become symptomatic after stopping long-term HRT or how best to discontinue therapy. Characteristics that might distinguish women who will have difficulty stopping therapy, such as severity of the original symptoms, body mass index, duration of hormone use, among others, have not been defined.

While many women have no difficulty stopping HRT, others like Mrs W develop vasomotor symptoms soon after discontinuation. In most women, these symptoms are mild and resolve over a few months. However, in a small subgroup of women, perhaps 1 in 10, recurrent symptoms are severe and persistent. With little published evidence to guide management of women like Mrs W, consultation with colleagues around the country suggests that most favor tapering therapy in women who have difficulty quitting, and most use either a "dose taper" or a "day taper." The dose taper involves progressively decreasing the dose of estrogen, for example, by dropping from 0.625 mg/d of conjugated estrogen to 0.4 or 0.3 mg/d, then discontinuing therapy. Mrs W tried this approach without success, but perhaps she attempted to taper too quickly. If a lower dose is associated with recurrence of symptoms, the next reduction in dose should not occur until symptoms improve and are easily tolerable, which may require 3 to 6 months in women who have difficulty. The day taper involves decreasing the number of days per week of HRT use, effectively decreasing the weekly dose. For example, therapy with the same dose of HRT may be continued, but only Monday through Friday. If this reduced weekly dose is tolerated, therapy might be discontinued on Thursday, and so on. As with the dose taper, if symptoms develop the weekly dose should be maintained until improvement occurs. Many clinicians combine the dose and day tapers by first halving the dose, then decreasing days of use. The day taper has the theoretical advantages of providing smaller reductions in weekly dose and ensuring that the estrogen dose is appropriately opposed by a progestin in women using combined continuous therapy. However, these approaches have not been studied individually or compared.

For most women who cannot tolerate even a slow HRT taper, the value of symptom relief probably outweighs the increased risk of venous thrombosis and breast cancer. However, for women at high risk of either of these conditions, like Mrs W with her relatively high risk of breast cancer, it might be reasonable to try an alternative agent for treatment of symptoms during the taper, such as venlafaxine,¹⁰ paroxetine,¹¹ or clonidine.^{9 ,12} Many women with symptoms turn to alternative therapies such as soy protein or herbal medications for relief of flushes. However, randomized trials of these products have generally found little benefit over placebo treatment.^{59 - 60}

Women like Mrs W who are at high risk of breast cancer might consider using either tamoxifen or raloxifene to reduce this risk. Recent large randomized trials of these drugs suggest that both substantially reduce the risk of breast cancer (Table 3). Among 13 388 women at high risk for breast cancer, 4 years of treatment with tamoxifen resulted in a 50% reduction in risk of both invasive and noninvasive breast cancer.¹⁶ However, risk was reduced only for estrogen receptor–positive breast cancer. Similarly, a large trial of the effect of raloxifene on risk for fracture among 7705 women with osteoporosis showed a 65% reduction in risk of invasive and noninvasive breast cancer associated with 3 years of treatment.⁶¹ Again, the benefit was limited to estrogen receptor–positive cancers. The absolute benefit of treatment depends on the underlying risk for breast cancer of women who are treated. For example, even though the relative reduction in risk for breast cancer was somewhat less for tamoxifen than raloxifene, the absolute benefit was much higher among women in the tamoxifen trial (absolute risk reduction, 21.4 cases of breast cancer per 1000 women treated vs 2.7 per 1000 with raloxifene) because the underlying risk of breast cancer was much higher among women in the tamoxifen trial (43.4 vs 3.6 cases per 1000 in the placebo groups).^{16 ,61}

While these drugs provide an exciting opportunity to reduce breast cancer risk, each has potential adverse effects that must be weighed against the potential benefit. Tamoxifen substantially increases risk of uterine cancer,^{16 ,62} but raloxifene does not.⁶¹ Like estrogen, both tamoxifen and raloxifene increase risk of venous thrombosis about 3-fold.^{16 ,50 ,62} Using population-based estimates of the incidence of venous thromboembolic events,⁶³ an increase of about 2 events per 1000 treated women aged 50 to 59 would be expected, and an increase of about 4 events per 1000 women aged 60 to 69. Clear guidelines for the use of tamoxifen and raloxifene for preventive therapy have not been established, but many experts suggest that women who are at high risk, like Mrs W with more than a 3% to 5% risk of breast cancer in the next 5 years, will likely benefit. For women who have undergone hysterectomy, I prefer prescribing tamoxifen over raloxifene because its efficacy was established in a trial among women at high risk for breast cancer, there is extensive clinical experience with this older drug, and it is the only drug approved by the FDA for this use. In women who have not had a hysterectomy, I generally prefer prescribing raloxifene to avoid the increased risk of uterine cancer associated with tamoxifen. Unfortunately, both drugs can cause or exacerbate hot flushes, currently Mrs W's major problem. If Mrs W had tapered estrogen without problems, I would consider treating her with tamoxifen to reduce her risk of breast cancer. But given that she still has severe hot flushes, I would not recommend tamoxifen at this time.

Many questions remain regarding the optimum use of tamoxifen and raloxifene. Physicians would like to treat only women who are at high risk for estrogen receptor–positive breast cancer, but currently there is no good way to identify these women. While 5 years of adjuvant treatment with tamoxifen reduced the risk of recurrence among women with early-stage breast cancer,⁶⁴ continuing therapy for an additional 5 years was associated with an 80% higher risk of recurrence compared with stopping.⁶⁵ The reason that risk for breast cancer recurrence may increase after long-term tamoxifen treatment is not clear, but it may be related to selection for resistant tumor cells or a change from initial inhibition to later stimulation of malignant cells.⁶⁶ It is not clear if a loss of benefit will occur in women without prior breast cancer or with long-term raloxifene treatment. For these reasons, the use of tamoxifen for the prevention of breast cancer is approved by the FDA for only 5 years. Even though treatment for longer than 5 years is not associated with additional risk reduction, the benefit of 5 years of tamoxifen use among women with breast cancer appears to persist for at least 10 years.⁶⁷ Because the randomized trials of tamoxifen and raloxifene in women without breast cancer continued for only 3 to 4 years, it is not clear how long the reduced risk of breast cancer observed in these women will persist.

Although there are options for relief of vasomotor symptoms, HRT remains the most effective therapy. When used for postmenopausal symptoms, HRT should be tapered and discontinued within about 5 years to minimize adverse effects. In any woman at high risk for osteoporotic fracture, bisphosphonates should be preferred to estrogen because the evidence of effectiveness is stronger and their risk profile is better. For prevention of disease, HRT should be restricted to use in women at high risk for osteoporotic fractures who are not at high risk for breast cancer. For postmenopausal women at high risk for breast cancer, HRT should be avoided, and treatment with tamoxifen or raloxifene to reduce risk of breast cancer should be considered.

Given Mrs W's risk of breast cancer and near normal bone density, I would recommend that she take calcium and vitamin D and try to stop taking estrogen. Although in the past she had severe symptoms when tapering her hormone therapy, she and her primary care physician can try again, but over an extended period. During that time, Dr K can consider adding a selective serotonin reuptake inhibitor, which may also help with symptoms. By tapering slowly, most women can discontinue HRT. However, some women—Mrs W might be one—cannot. In that case, she should continue taking hormone treatment for symptoms but not for disease prevention.

A PHYSICIAN: If alendronate is your first choice for treating a patient who is truly osteoporotic, is there additional benefit to adding estrogen in combination with the alendronate?

DR GRADY: The effect of alendronate plus estrogen on bone density has been examined but not the effect of their combination on fracture risk. In a 2-year clinical trial, alendronate alone and estrogen alone resulted in an equivalent increase in bone density compared with placebo. The combination of the 2 drugs was better than either alone, but the benefit was small—about 2% higher bone density in women using the combination compared with either drug alone.⁶⁸ This small increase in bone density is unlikely to translate into a clinically meaningful reduction in risk of fracture or to outweigh the potential adverse effects of adding estrogen to a bisphosphonate.

A PHYSICIAN: Could you comment on the risk of breast cancer mortality with estrogen use?

DR GRADY: While a large number of studies have examined the effect of HRT on risk for developing breast cancer, few have been large or long enough to examine risk of breast cancer mortality. Of those that have, the results are mixed. Some show no effect of HRT on breast cancer mortality,⁶⁹ others show an increased risk,⁷⁰ and 1 even shows a decreased risk.⁷¹ So the evidence regarding the effect of estrogen on breast cancer mortality is inconclusive. Still, I think most women want to avoid developing breast cancer, even if their risk of dying of it is not increased.