Since July 2002, the major news dominating headlines in medical publications dealing with women’s health has been the data from the WHI. Alarmingly, there is little consensus among both experts and practicing clinicians alike on how best to interpret the WHI results in this cohort of older women who were given estrogen-plus-progestin therapy. The general assumption of the lay press has been that randomized clinical trials are without study-design or interpretation flaws and therefore are the cornerstone of clinical practice. This unfortunate assumption fails to recognize the specific limitations of clinical trials–the very limitations that hinder this study.
As clinicians we are keenly aware of the WHI’s shortcomings. For instance, there have been concerns regarding the cohort studied. The WHI subjects were older than the typical menopausal patient, and generally without symptoms related to loss of estrogen. The cohort also represented a wide range of ages (from 50 to 70 years), and many patients with a markedly increased body mass index. Add to this some of the obvious idiosyncrasies of the data itself–for instance, the low rate of adverse events tracked in the placebo group for year-5 data–and the findings generate many questions for practicing clinicians and their patients.
Questions about this newly released data focus on the same issues raised by the original article, as well as a few new ones. The most obvious concern is a practical one that usually stirs no debate: How can a clinical trial study the effects of pharmacologic interventions on quality of life in a group of women who were not recruited because of poor lifestyle or dissatisfaction with their everyday living? In fact, women were excluded from this study if they reported symptoms related to the menopause.
Studying asymptomatic menopausal women is akin to prescribing antifungal or placebo vaginal preparation in a blinded manner to subjects without a yeast vaginitis, and then collecting data to evaluate which group is most satisfied after intervention.
Prior studies using hormonal therapy in women with symptoms related to hypoestrogenism have shown improvement in vasomotor symptoms and quality of life.1,2 Women without symptoms related to estrogen loss cannot be expected to show an improvement in well-being with the use of hormone therapy.
My other concern is that these data may spur an increased use of unproven alternative therapies. For example, most nonhormonal interventions, such as herbal and dietary supplements, have either scant or nonexistent data supporting their role in improving patient well-being.
The question for clinicians is obvious: Is there a place for hormonal therapy in improving quality of life? To practice the bestquality medicine, we must use the data generated by randomized clinical trials, observational trials, retrospective reviews, and clinical judgement.
Each has a role in forming opinion: A randomized controlled trial recruits women willing to accept placebo intervention, randomization to different treatments, and a certain degree of the unknown.
Data from observational trials complement randomized controlled trial data because they more accurately reflect standards of medical care and either refute or support existing management protocols.
Retrospective reviews offer useful information about trends and prove helpful as pilot data to construct prospective trials.
Lastly, clinician acumen and the doctor-patient relationship take into account the individual needs of women. The best clinical practice does not rely on 1 set of data derived from a single data set.
The data from this article should be shared with women, but should not overshadow other data on hormone therapy or the individual needs of an informed patient. Rather, it should be a piece of the pie–i.e., the aggregation of all objective data on HRT–and assist the woman and her physician in deciding whether the benefits outweigh the risks for her specific situation.