HT: a clinician demurs☆

Cardiovascular disease 

Are the small increases in cardiovascular events reported by the WHI real or are there other explanations for the reported results? This is an important question because the reported increases were not large; a change in a small number of cases could alter the WHI conclusions. It is noteworthy, for instance, that a prospective, nested case-control study from the WHI reported that with comparable levels of C-reactive protein (CRP), the use of hormone therapy did not change the odds ratios for coronary heart disease, despite the increase in CRP that occurs with hormone therapy (1).

It is hoped that subsequent analyses of the WHI data will address the following concerns: In the Heart and Estrogen/progestin Replacement Study (HERS), statins and aspirin protected against venous thromboembolism, and most importantly, the reported increase in arterial events in the first year of the trial was no longer significant when statin and aspirin treatment was considered (2). A higher prevalence of new statin and aspirin treatment in the placebo arm of the WHI could have produced lower event rates, giving the false impression of higher rates in the treatment arm—this is a critical analysis yet to be provided.

The published results depended on cardiovascular diagnoses made “in the field.” In the updated report from the WHI, adjudication of the diagnoses produced a 10% disagreement for myocardial infarction and 3% for death due to coronary heart disease, and this small level of disagreement changed the strength of the conclusions (indeed, the overall results by definition now do not achieve statistical significance) (3). The possibility of diagnostic bias should also be considered. In the WHI, 40.5% of the treated group, in contrast to 6.8% of the placebo group, was unblinded because of vaginal bleeding. What would be the impact on the clinician's final management and diagnosis of cardiac symptoms when told that the patient was in the WHI study and experiencing vaginal bleeding?

Will postmenopausal hormone therapy begun at or near the time of the menopause, and maintained for a relatively long duration of time, provide protection against coronary artery disease? The design of the canceled arm of the WHI did not allow an answer to this question. Women with significant menopausal symptoms were excluded from the study to avoid an exceedingly high dropout rate in the placebo group, resulting in a study population with an average age of 63 and 18 years postmenopause. The WHI investigators addressed this problem by pointing out that the ratios of cardiovascular events in the treated and placebo arms were the same when assessed by decades of age: 50s, 60s, and 70s. However, this is not the vital analysis. By excluding women with menopausal symptoms, it is likely that only a small number of the participants were close to their age of menopause. In the updated report, only the women who were 20 or more years distant from menopause had a statistically significant increased risk of coronary heart disease. This emphasizes the secondary prevention nature of the WHI.

Breast cancer 

For nearly a decade, we have been teaching that the lack of a uniform, consistent conclusion in more than 60 case-control and cohort studies on breast cancer and postmenopausal hormone therapy means that any effect has to be a small one. The WHI results do not change this. The most important unanswered question is whether postmenopausal hormone therapy initiates the growth of new breast cancers or whether the epidemiologic results reflect an impact on pre-existing tumors. Observations that favor an impact on pre-existing tumors include: [1] the return of the hazard risk in the WHI study almost to 1.0 in year 6, [2] no difference in noninvasive breast cancers in the WHI treatment and placebo arms, and [3] the large body of literature documenting lower grade and stage disease in hormone users, resulting in better survival rates. The evidence is consistent with earlier detection of breast tumors in hormone users, either because pre-existing tumors are stimulated to grow faster and appear sooner or because hormone-induced differentiation actually slows tumor growth, allowing more time for early detection. The WHI did agree with convincing evidence that postmenopausal hormone therapy does not increase the risk of breast cancer beyond that already associated with recognized risk factors, such as a positive family history.

The WHI updated report on breast cancer and the report from the UK Million Women Study disagree with a large body of literature indicating that hormone users who develop breast cancer have smaller and better differentiated tumors with better survival rates 4, 5. It is not appropriate to assume that new studies are right and old studies are wrong. The challenge is to identify the reasons for the disagreements among the studies. The discovery and understanding of those reasons will reduce the current polarization between proponents and opponents of hormone therapy, and help to establish the proper use of this treatment for individual women.

Clinicians must continue to exercise medical judgment, making informed decisions to translate the accumulated knowledge into effective and appropriate clinical practice.

Clinical Conclusions 

Clinicians must continue to exercise medical judgment, drawing on the large body of knowledge accumulated over the past 20 years, making informed decisions to translate the accumulated knowledge into effective and appropriate clinical practice. My own clinical judgments, thus far, lead me to the following conclusions: