June 2002 I read with interest the article by Dr Montemuro on treatments for menopausal symptoms [BCMJ 2001;43(8):452-457], but I cannot agree with her assessment of the evidence in favor of the clinical use of St. John’s wort, black cohosh, ginkgo, or valerian.
Two key elements in the critical evaluation of any therapeutic agent are the plausibility of the supportive evidence and the assessment of potential harm from its use. Convincing evidence of safety and effectiveness does not exist for any of the four agents listed above.
In the case of black cohosh the most recent (and only) RCT for menopausal symptoms found it to be no better than placebo.[1]
The three English-language RCTs on valerian paint a murky picture, with the most rigorous tests showing no effect:
Donath and colleagues[2] reported multiple endpoints in a study of valerian extract on sleep structure. While some of the endpoints showed statistical significance in favor of valerian, the clinical significance is far from clear. A more powerful finding would have been clear statistical and clinical significance demonstrated on an a priori endpoint.
Schulz and colleagues[3] also measured multiple endpoints when comparing placebo to valerian. While positive effects were noted for slow wave sleep, no effect was seen on subjective sleep quality, time to sleep onset, or REM sleep.
Balderer[4] reported a mild sedative effect of valerian over placebo, however the trial numbers were small (n=18) and the part of the trial conducted under the most rigorous conditions showed no significant effect.
The case of St. John’s wort is instructive in that it illustrates how poorly done studies may give the appearance of a credible body of evidence without actually comprising one. As Dr Montemuro points out, there are over 20 RCTs supportive of St. John’s wort’s effectiveness in the treatment of depression. What critical appraisers will find disturbing, however, is the lack of long-term safety data, the generally poor case definitions, the small numbers involved, the failure to control for duration of illness, and the relative lack of replication of these studies by people other than herbal proponents. Independent reviewers have been concerned about the quality of the so-called “body of evidence” behind St. John’s wort, and excellent discussions of the experimental design flaws seen in these papers can be found in Deltito[5] and Field.[6] Of even greater concern is the fact that the one study that employed the greatest scientific rigor found there to be no efficacy greater than placebo.[7]
The literature on Ginkgo biloba paints a picture most consistent with a remedy in search of a condition to treat. Although ginkgo is alleged to be of use for ADHD,[8] depression,[9][10] poor sleep,[11] and sexual libido,[12] the relevant studies are either negative (tinnitus, depression) or of such poor quality as to make conclusions about efficacy impossible (poor sleep, sexual libido). tinnitus,
For example, the RCT by Hemmeter and colleagues[11] on sleep employed neither blinding nor controls. The study by Waynberg[12] reporting increased libido in pre- and post-menopausal women was conducted without the benefit of randomization, blinding, or a placebo group.
Presumably, the rationale for the use of ginkgo in menopause would be to address symptoms of depression, poor sleep, and decreased libido. Even if one were to consider the RCTs in which ginkgo is said to improve memory loss,[13-21] the quality of the studies leaves much to be desired. In general, the studies suffer from low numbers (13, 16, 18, and 20 to 22), or the use of multiple endpoints (13, 17, and 22). Three of the studies report discernable small improvements on certain cognitive scoring scales but do not establish the clinical significance of these results (14, 15, 17, and 19). Also weighing against these papers are the essentially negative studies of Moulton[23] and Mix[22] and the large RCT by van Dongen,[24] which found no effect.
Similarly, the evidence of the cardioprotective effect of a low-fat diet, far from being difficult to overemphasize, is quite tenuous and indicates that a strict low-fat diet could at best add 2 months to one’s life expectancy.[25]
Dr Montemuro asserts that “Women at high risk for breast cancer will need to avoid hormones and use lifestyle modification and complementary and alternative therapies almost exclusively.” This statement fails on two counts:
The therapies advocated need rigorous proof before being used in clinical practice.
It does not make sense to assume that biologically active herbal remedies can have only good long-term effects if no studies have been done specifically looking for the undesirable endpoints that are alleged to be the result of conventional therapy.
Patients with chronic or incurable conditions are especially attracted to unproven treatments. Physicians need to know not only what supportive evidence exists for these therapies, but also how to rate its quality.
—Lloyd Oppel, MD
Vancouver
Thank you for inviting me to respond to Dr Oppel’s letter regarding the article I wrote, “Relieving the symptoms of menopause: From herbs to hormones.”
I based the article on information gathered at the SOGC (Society of Obstetricians and Gynaecologists of Canada) Consensus on Menopause and Osteoporosis, the meetings of which I attended during the winter/spring 2001. The members of the SOGC Menopause and Osteoporosis Consensus included gynecologists, internists, endocrinologists, a nurse, a pharmacologist, a family physician, a medical researcher, and members of the public. I also referred to the Canadian Medical Association reference book Herbs: Everyday Reference for Health Professionals.
During the SOGC meetings we examined the quality of evidence for all therapeutic interventions including lifestyle, vitamins, hormones, antidepressants, oral contraceptives, osteoporosis medication, and herbal remedies. The members of the Consensus assigned levels of evidence to the research we examined. The proceeding of these meetings were printed in the SOGC’s Journal of Obstetrics and Gynaecology Canada September-December 2001. I have referenced some of the chapters from the Consensus below.
I certainly share Dr Oppel’s concern about the widespread and often inappropriate use of herbal remedies for which there is scant scientific data. The purpose of including information about herbal as well as traditional therapies in this article was to inform physicians about a select few herbs, including warnings about their interactions and side effects. These select herbs have randomized control trials, although of short duration, therefore short-term use was advised. By reviewing the relevant data on a select number of herbs I hoped to give physicians some useful tools to help them answer women’s questions about alternative management of menopausal symptoms.
I am concerned that Dr Oppel felt there was not enough scientific data to recommend a low-fat diet for cardioprotection. Is he suggesting we stop trying to convince our patients to improve their diets? The Nurses Health Study showed an 83% reduction in coronary events in women who adhered to a healthy lifestyle (low-fat diet, exercise, and non-smoking). Unfortunately only 3% of the nurses actually fit the criteria for a healthy lifestyle. This finding underscores the enormous potential for intervention in this area.
I refer all readers with questions to the Consensus on Menopause and Osteoporosis.
—Suzanne Montemuro, MD
North Vancouver
Suggested reading
Chandler F (ed). Herbs: Everyday Reference for Health Professionals. Ottawa, ON: Canadian Medical Association and Canadian Pharmacists Association, 2000.
Derzko C, Fluker M, Montemuro S, et al. Menopause: Healthy Living: Canadian Consensus on Menopause and Osteoporosis. J Obstet Gynecol Can 2001;23:14-20. Full Text
Fluker M, Montemuro S. Complementary Approaches: Canadian Consensus on Menopause and Osteoporosis. J Obstet Gynecol Can 2001;23:70-79. Full Text
Stampfer M, Hu F, Manson J, et al. Primary prevention of coronary heart disease in women through diet and lifestyle. N Engl J Med 2000;343:16-22.PubMed Abstract Full Text