Hormone replacement therapy and female malignancy: What has the Million Women Study added to our knowledge?

Abstract

The trends in breast, endometrial and ovarian cancer incidence reported by the MWS are broadly similar to previous and subsequently published clinical evidence. The cancer incidence data, when viewed in absolute terms in comparison with the WHI study, support this (Table 3). Although the MWS findings are based on a very large cohort, these are observational data and should be interpreted in the context of other evidence, particularly randomised controlled trials. If differences exist between women attending the NHSBSP and those who do not, and between attendees who agreed or declined to participate in the study, this could bias the reported results. Women who use HRT are also more likely to exhibit health-orientated behaviour than non-users and may be offered more health checks in primary care. These differences cannot be accounted for by the MWS design but are important in its interpretation. The size of this study alone is insufficient reason for its findings to be accepted without question. The mortality data are impossible to interpret with accuracy for reasons outlined previously. If anything, it is the interpretation and emphasis placed on their findings by the MWS investigators that has created confusion and concern. Counselling women about HRT is complex, due to its diverse profile in relation to cancer- and non-cancer-associated outcomes. The MWS investigators constantly stress the former. Despite the fact that all their publications emphasise that individual findings on breast, ovarian and endometrial cancer should not be viewed in isolation, they present their data in complete absence of other clinical evidence on chronic health conditions affected by HRT that have just, if not more, potential for impacting on a woman's overall survival. These include colorectal cancer (where combined therapy has a protective effect), cardiovascular disease (where there is likely to be benefit if HRT is commenced in the perimenopause or within 10 years of the menopause), osteoporosis (again a beneficial impact) and thromboembolic disease (where risk is initially increased). Their arguments, in addition, do not address the issue of symptom relief and quality of life benefits with HRT and are prescriptive to say the least. Any practitioner discussing HRT is (or should be) aware of the complexities in counselling women adequately. Discussion of known risks and benefits, including current areas of uncertainty, would seem to be a more measured and rounded means of providing public health education, but it probably wouldn't stimulate the production of press releases and generate media headlines. This is a shame, as it stifles debate and ignores the fact that most women are pragmatists and quite capable of weighing up health information, even if this information is uncertain. HRT is still the most effective treatment for the relief of oestrogen deficiency symptoms, and for most women who commence HRT for this indication the relatively short-term exposure (i.e. <5 years) is unlikely to have an adverse impact on their overall health whilst improving their well-being. ©FFPRHC 2007.