The WHI through the Restrospectoscope

There is a concise, well-written discussion of the findings of the Women’s Health Initiative in this week’s Jama by Steven Hully and Deborah Grady. It’s the best short summary of the issues I have read to date, and I would encourage you to read it. I find that I agree with almost all of it.

As a practicing gynecologist caring for women in menopause, however, I find myseld a little more critical of the original study design, which enrolled primarily asymptomatic women on average 10 years post menopausal, which is not how HRT is generally used.

A residual loose end stems from the fact that the average age of women at enrollment for all 3 trials was the mid-60s, whereas the majority of women who use hormone therapy for treatment of vasomotor symptoms are in their early 50s. Hormone therapy among women aged 50 through 59 years in the WHI trials tended to have more favorable effects on CHD rates than in older women, but even if this finding is confirmed, the concerns about increased risk of stroke, venous thromboembolism, and breast or endometrial cancer remain.

This is a lot more than a “residual loose end” – it is a major flaw in the study design that understandably resulted from the good intentions of the investigators not to do harm by randomizing symptomatic women suffering from hot flashes to a placebo. I would agree that in younger women there remains a concern about thromboembolism and stroke, but the use of transdermal lower dose preparations may mitigate much of this risk. (Notice I say “may” – we still don’t have the clinical data we need.) The other big “loose end” is the issue of bioidentical hormones vs horse estrogens and medroxyprogesterone acetate, the drugs used in this study. Both these issues will require another large randomized clinical trial before we can say we have done this issue justice for menopausal women. Fortunately, such studies are now being conducted, and hopefully we will have some of the answers we need in a few more years.

I’m not sure why the risk for endometrial cancer are mentioned in the editorial – it did not appear in the WHI results, and should not be an issue if progesterone is taken according to recommended HRT protocols.

Still, the WHI answered many important questions about HRT. In particular, the decline in breast cancers that has resulted since hormone use dropped after the publication of its results more than supports this trial as one of the most important in the history of American healthcare.

In conclusion, the story of HERS and WHI is an excellent illustration of the evidence-based medicine tenet that practice guidelines should be based on rigorously designed research—preferably 2 or more randomized blinded trials with disease end points—even if consistent observational and mechanistic evidence suggests that such trials are not needed. Animal studies and clinical trials of surrogate outcomes can be misleading, and epidemiologic studies of preventive treatments are particularly susceptible to confounding because healthier individuals are more likely to seek and adhere to preventive measures. Weighing benefits and harms is especially important when considering the use of preventive interventions in healthy individuals, in whom there is a special obligation to do no harm. (italics mine)

It is so important to remember that before the WHI, we were prescribing hormone replacement to women just because they were menopausal. When a drug is being used in healthy individuals, the bar is raised to its highest in terms of safety.

Primum non nocere.