BENEFICIAL OR DANGEROUS?
The Women’s Health Initiative (WHI) conducted by the National Heart, Lung and Blood Institute of the NIH prematurely terminated a portion of a large study on postmenopausal hormone replacement therapy (HRT) after finding a 26% increased risk of breast cancer among those taking the active medication as compared to those taking placebo. The announcement, which accompanied an article in the Journal of the American Medical Association on July 27th of 2002, has been widely publicized in the media as proof that hormone replacement therapy is dangerous.
Some television viewers were told by media personalities to immediately stop taking all such medications. And many women did. Many more were needlessly alarmed and confused. The news was a windfall opportunity for peddlers of "alternative" menopausal symptom remedies, who immediately began to further misrepresent the WHI findings in order to boost sales of soy-based, "phytoestrogen," and various herbal remedies. Even the reputable GlaxoSmithKline, manufacturers of the black-cohosh "nutritional supplement" Remifemin®, rushed into sponsoring a "throwaway" medical journal "Special Edition on Alternative Therapy for Menopause."
The facts about the WHI are somewhat different:
- Only the portion of the WHI in which the women were taking a particular oral combination of one kind of estrogen (conjugated equine estrogens or CEE) and one kind of progestin (medroxyprogesterone acetate or MPA) on a continuous basis was discontinued. MPA is known to exert adverse effects on serum lipids that are not seen with other progestins. Meanwhile, the WHI’s estrogen-only arm – also using CEE – is continuing because no excess risk was found.
- The study subjects ranged in age from 50-79 with the average age being 63. That is, most women in the WHI had already been without estrogen for many years and, in the real world, few such women are started on HRT.
- Substantial numbers of WHI’s subjects dropped out or switched from the treatment to the placebo group or vice-versa. By the end of the study this amounted to well over a third of the women being studied. This represents a serious degradation of the quality of the data.
- The 26% increased risk of breast cancer is in the range of findings from previous studies and is very small in absolute terms. And, as in previous studies, no increased mortality from breast cancer was found.
- Since it typically takes 5-7 years for a breast cancer to be detected once it begins, the “increased” cases of breast cancer found by WHI were a result of estrogen’s promoter effect and not from actual carcinogenesis. Indeed, it is known that women diagnosed with breast cancer who are on HRT have a better prognosis. This too fits with what is known of the biology of breast cancer: that the process of transformation to malignancy takes time and that the most aggressive tumors are those that have had both time and opportunity to lose their sensitivity to hormones.
- As previous research found, the WHI also demonstrated reduced risks of colon and rectal cancer, uterine cancer and hip fracture among HRT users. The reduction in colon and rectal cancer is of particular importance because these malignancies carry a much worse prognosis than that of breast cancer.
- The WHI did not adequately consider many other benefits of HRT including relief of hot flushes and vaginal atrophy, sexual function, cognitive function and dentition. Additional information from the WHI purports to show that HRT does not offer these benefits, which is suspicious given the body of evidence already accumulated that clearly show that it does. This discrepancy may be explained, again, by the fact that most of the WHI study subjects were not newly menopausal and that many women crossed over from the treatment to the placebo arm and vice versa.
- Although the randomization of the study participants appears to have been effective with respect to most variables, no information was presented with respect to known risk factors for breast cancer such as age at menarche, age at first pregnancy/childbirth, breastfeeding, age at menopause, and presence or absence of the BRCA1 and BRCA2 genes.
The year-by-year findings (Table 4 in the JAMA article) are particular revealing with respect to the ratios of events in the HRT group compared to placebo: (1 is normal risk, less than 1 is reduced risk and greater than 1 is increased risk - but keep in mind that the absolute risks are small. That is, if normal risk is .1% then a hazard ratio of 2 is still only a .2% risk.)
|
YEAR
|
1
|
2
|
3
|
4
|
5
|
6
|
| Coronary Disease |
1.78
|
1.15
|
1.06
|
.99 | 2.38 | .78 |
| Stroke |
.95
|
1.72
|
1.79
|
1.70
|
1.87
|
.66
|
| Venous Thromboembolism |
3.60
|
2.26
|
1.67
|
1.84
|
2.49
|
.90
|
| Invasive Breast Cancer |
.62
|
.83
|
1.16
|
1.73
|
2.64
|
.90
|
| Endometrial (Uterine) Ca |
.95
|
.96
|
.76
|
1.91
|
.23
|
.17
|
| Colorectal Cancer |
.64
|
1.17
|
.72
|
.43
|
.47
|
.59
|
| Hip Fracture |
.64
|
.59
|
.87
|
.69
|
.58
|
.55
|
| Total Mortality |
1.24
|
.96
|
1.06
|
1.09
|
.87
|
.83
|
As in the HERS study which looked at HRT for secondary prevention of cardiovascular disease, most of the increased risk from cardiovascular events is in the first few years. Otherwise the year-to-year figures are most consistent with respect to reductions in colorectal cancer and hip fracture. Indeed, these were the only two outcomes measured in which the 95% confidence interval did not include unity (that is, a hazard ratio of 1, no different than placebo.)
The breast cancer numbers are entirely consistent with the idea that HRT can promote pre-existing subclinical disease. This is also borne out by the lack of any increased mortality from breast cancer (or from any other cause), and it may very well be that pre-existing subclinical breast cancers, allowed to progress more slowly over a longer period of time among women not taking HRT, could eventually result in more advanced cancers of higher grade. The breast cancer "bump" in year 5 of the WHI study is particularly inOnly the portion of the WHI in which the women were taking a particular oral combination of one kind of estrogen (conjugated equine estrogens or CEE) and one kind of progestin (medroxyprogesterone acetate or MPA) on a continuous basis was discontinued. MPA is known to exert adverse effects on serum lipids that are not seen with other progestins. Meanwhile, the WHI’s estrogen-only arm – also using CEE – is continuing because no excess risk was found.
The study subjects ranged in age from 50-79 with the average age being 63. That is, most women in the WHI had already been without estrogen for many years and, in the real world, few such women are started on HRT.
Substantial numbers of WHI’s subjects dropped out or switched from the treatment to the placebo group or vice-versa. By the end of the study this amounted to well over a third of the women being studied. This represents a serious degradation of the quality of the data.
The 26% increased risk of breast cancer is in the range of findings from previous studies and is very small in absolute terms. And, as in previous studies, no increased mortality from breast cancer was found.
Since it typically takes 5-7 years for a breast cancer to be detected once it begins, the “increased” cases of breast cancer found by WHI were a result of estrogen’s promoter effect and not from actual carcinogenesis. Indeed, it is known that women diagnosed with breast cancer who are on HRT have a better prognosis. This too fits with what is known of the biology of breast cancer: that the process of transformation to malignancy takes time and that the most aggressive tumors are those that have had both time and opportunity to lose their sensitivity to hormones.
As previous research found, the WHI also demonstrated reduced risks of colon and rectal cancer, uterine cancer and hip fracture among HRT users. The reduction in colon and rectal cancer is of particular importance because these malignancies carry a much worse prognosis than that of breast cancer.
The WHI did not adequately consider many other benefits of HRT including relief of hot flushes and vaginal atrophy, sexual function, cognitive function and dentition. Additional information from the WHI purports to show that HRT does not offer these benefits, which is suspicious given the body of evidence already accumulated that clearly show that it does. This discrepancy may be explained, again, by the fact that most of the WHI study subjects were not newly menopausal and that many women crossed over from the treatment to the placebo arm and vice versa.
Although the randomization of the study participants appears to have been effective with respect to most variables, no information was presented with respect to known risk factors for breast cancer such as age at menarche, age at first pregnancy/childbirth, breastfeeding, age at menopause, and presence or absence of the BRCA1 and BRCA2 genes.
teresting. Without it, it seems doubtful that the overall 26% increased risk would have been reported.
Thus, the principal findings of the WHI study reported in July of 2002 was that primary prevention of cardiovascular disease by this particular combination of HRT among already postmenopausal women who have not been on HRT is ineffective. The findings with respect to breast cancer broke no new ground and, in fact, leave it open to question as to the relative role and importance in the genesis of breast cancer of the tiny doses of hormones used in HRT as compared to the much higher amounts of these substances in a woman’s body during her reproductive years. It may be that surgical removal of the ovaries as soon as a woman has completed her childbearing would be an effective way to reduce the incidence of breast cancer. This is known to be the case among women at high risk of breast cancer. Alternatively, medical suppression of ovarian function unless conception is desired could have similar beneficial effects.
But would men agree to undergo castration once they had fathered their children? Could they be induced to forgo androgen replacement therapy when it would otherwise be appropriate with the idea that it would reduce their risk of prostate cancer? As it is, men are being advised in advertising that if they are tired and "don’t feel like the man you used to be" that they should take various "dietary supplements" to increase their levels of testosterone. Of course, other ads tell them to take different preparations in order to cure their prostate troubles or to reduce the "bad" androgenic hormone implicated in baldness. These issues involve philosophical questions more than medical ones.
Dr. Wolf Utian, Executive Driector of the North American Menopause Society, recently made these observations about medical research studies generally and the WHI study in particular:
“The discussion and conclusion [of medical scientific research reports] may expose the authors’ bias ... It is in this section that truth most often becomes opinion, and conclusions or recommendations may not be justified by data. The media typically pick up the opinions rather than the more complicated data. In this way, scanning the abstract and conclusion may result in disinformation to the public at large. Excellent examples of this disconnect exist in the medical literature on Women’s Health Initiative studies into hormone usage. ... overinterpretation of the WHI results and their incorrect transposition to a younger, perimenopausal population have had a major impact on clinical practice. This raises the question, ‘When does “statistically significant” equate to “clinically relevant?”’ There is no simple answer. The WHI publications described breast cancer findings as being statistically insignificant but clinically relevant, but reported certain quality-of-life parameters as statistically significant but, in the investigators’ opinion, clinically irrelevant.” [ in “Lessons from the WHI: Randomized, Controlled Trials and The Need To Read The Details,” Ob/Gyn Perspectives (Cleveland Clinic), September 2005, p 10-11.]
The overall conclusion for now concerning the WHI is that there is still no better treatment for menopausal symptoms than estrogen, of which there are several varieties besides CEE, all of which offer benefits in prevention of osteoporosis and colorectal cancer. It remains an open question as to whether these medications may offer primary prevention benefits for cardiovascular disease if begun as soon as menopause occurs. For women with a uterus, a progestin is also indicated, but this is probably best not given as continuous MPA. Rather, periodic MPA withdrawal or continuous or periodic use of another progestin, of which there are several, should be preferred. In addition, there is now an extended-use progestin-containing IUD (Mirena®) that can be used. Women who are already taking continuous CEE-MPA and experiencing no problems can be reassured or, if they prefer, switch to another regimen.
A woman on HRT who experiences an adverse cardiovascular event should be managed according to the specific cirumcstances based on the preparation she has been taking and for how long and on other, more important risk factors for recurrence. Every woman should be informed that her risk of dying from breast cancer is about 5%, which is related primarily to her being a woman, while her risk of death from cardiovascular disease is about 90%. Every woman should also be educated about the overwhelming importance of the most important cardiovascular risk factors which are smoking, sedentary lifestyle, elevated cholesterol and triglycerides, low HDL cholesterol and diabetes/insulin resistance.
It is irresponsible, on the basis of the WHI findings, to discontinue all patients on estrogen replacement, lead them to believe that such therapy is dangerous or that it has been proven that the risks outweigh the benefits, or that unproven and ineffective "alternatives" are a superior choice.