Breast cancer and hormone-replacement therapy: the Million Women Study
2003; Elsevier BV; Volume: 362; Issue: 9392 Linguagem: Inglês
10.1016/s0140-6736(03)14596-5
ISSN1474-547X
AutoresValerie Beral, Emily Banks, Gillian Reeves, Diana Bull,
Tópico(s)Menopause: Health Impacts and Treatments
ResumoAuthors' reply Sir–Several correspondents state that findings from the Million Women Study are inconsistent with, or may overestimate, HRT-associated risks of breast cancer when compared with results reported from the collaborative reanalysis of the worldwide data1Collaborative Group on Hormonal Factors in Breast CancerBreast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer.Lancet. 1997; 350: 1047-1059Summary Full Text Full Text PDF PubMed Scopus (2497) Google Scholar and the WHI trial.2Chlebowski RT Hendrix SL Langer RD et al.for the WHI InvestigatorsInfluence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women's Health Initiative randomised trial.JAMA. 2003; 289: 3243-3253Crossref PubMed Scopus (1717) Google Scholar In fact the results are remarkably similar. The worldwide collaboration estimated an excess incidence of breast cancer of two and six per 1000, respectively, for 5 and 10 years' use of HRT (80% of the HRT users used oestrogen-only preparations)1Collaborative Group on Hormonal Factors in Breast CancerBreast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer.Lancet. 1997; 350: 1047-1059Summary Full Text Full Text PDF PubMed Scopus (2497) Google Scholar, comparable to estimates of 1·5 and five per 1000 users, respectively, for 5 and 10 years' use of oestrogen-only HRT in our study. For combined oestrogen-progestagen HRT, the excess incidence of breast cancer in women who complied with treatment in the WHI trial is about six and 18 per 1000, respectively, for 5 and 7 years' use of HRT; comparable to estimates of six and 19 per 1000, respectively, for 5 and 10 years' use in our study. Findings that the increased risk of breast cancer in current users of HRT in the Million Women Study wore off a few years after use ceased are consistent with the worldwide data1Collaborative Group on Hormonal Factors in Breast CancerBreast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52 705 women with breast cancer and 108 411 women without breast cancer.Lancet. 1997; 350: 1047-1059Summary Full Text Full Text PDF PubMed Scopus (2497) Google Scholar Also, the observed increase in mortality is consistent with the more advanced stage of breast cancers in the HRT compared with the placebo group in the WHI trial2Chlebowski RT Hendrix SL Langer RD et al.for the WHI InvestigatorsInfluence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women's Health Initiative randomised trial.JAMA. 2003; 289: 3243-3253Crossref PubMed Scopus (1717) Google Scholar. We used standard methods for analysis of cohort data and the alternatives proposed are inappropriate and/or inapplicable in this setting. For example, the suggested survival analyses among women with breast cancer cannot answer the question we set out to address-whether use of HRT increases death rates for breast cancer. Further subdivision of results on breast cancer for tibolone users (included with other types of HRT in figure 3 in our original paper) would not give meaningful results at present, because the associated standard errors are too large. With respect to other effects of tibolone, the Committee on Safety of Medicines has advised health professionals that “the risk of endometrial cancer with tibolone is unknown”3Cmmittee on Safety of Medicines Hormone replacement therapy (HRT) and breast cancer: results of the UK Million Women Study.http://www.mhra.gov.ukGoogle Scholar. Correspondents suggested various hypothetical potential biases associated with recruiting a cohort at the time of routine invitation for mammography, and some requested separate results for screen-detected and other cancers. The interpretation of such results is complex. For example, the relative risk for current-users versus never-users of HRT at baseline is 1·4 (95% CI 1·3-1·5) for screen-detected breast cancers diagnosed immediately after recruitment, but 2·7 (2·2-3·3) for other breast cancers diagnosed within a year of recruitment (8 months later, on average); corresponding results for women with less than 1 year's use of HRT at baseline are 0·8 (0·6-1·1) and 3·8 (2·3-6·1), respectively. The results for screen-detected versus other cancers differ substantially (p<0·0001 for each). That these large differences are due to an extra 8 months' use of HRT or to the selective recruitment of women with breast problems seems unlikely. HRT reduces mammographic sensitivity4Banks E Hormone replacement therapy and the sensitivity and specificity of breast screening: a review.J Med Screen. 2001; 8: 29-34Crossref PubMed Scopus (69) Google Scholar which seems the most likely reason for the differences. We plan to publish detailed results on the relation between the use of HRT, the timing of breast cancer diagnosis in relation to mammography, and tumour characteristics. However, because HRT alters mammographic sensitivity, the most appropriate way to investigate the effect of HRT on the underlying risk of breast cancer is to combine screen-detected and other breast cancers, as was done. Many of the suggested mechanisms of action of HRT on the breast cannot be substantiated. The view that HRT acts solely by accelerating growth of pre-existent breast cancers is contradicted by the fact that no deficit is seen in past users. Also, as the investigators note, the WHI trial results indicate that the increased risk of breast cancer emerges sooner after the initiation of hormonal therapy than previous results had suggested2Chlebowski RT Hendrix SL Langer RD et al.for the WHI InvestigatorsInfluence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women's Health Initiative randomised trial.JAMA. 2003; 289: 3243-3253Crossref PubMed Scopus (1717) Google Scholar. After careful review of the correspondents' comments, we consider that none invalidates our main conclusions. Larger numbers of events are needed to answer some of their questions, and continued follow-up of the Million Women Study cohort should provide this additional information. Breast cancer and hormone-replacement therapy: the Million Women StudyThe Million Women Study (MWS) collaborators (Aug 9, p 419)1 report a doubling of the risk of incident invasive breast cancer in current-users of combined hormone-replacement therapy (HRT) compared with never-users, causing considerable alarm among patients, doctors, and the pharmaceutical industry.2 Media medics urged calm by emphasising absolute rather than relative risk, but for many women the damage was done. Unfortunately, this study could have overestimated the risk of breast cancer associated with HRT. Full-Text PDF Breast cancer and hormone-replacement therapy: the Million Women StudyThe MWS1 provides important information on the relative risk of combination oestrogen-progestagen HRT compared with other preparations. However, we are concerned that the study has over-estimated the risk of breast cancer associated with a short exposure to HRT. Full-Text PDF Breast cancer and hormone-replacement therapy: the Million Women StudyWe have some concerns about the results of the MWS,1 despite the collaborators' claims that the size of their study enables unbiased reporting of breast-cancer incidence and mortality rates. Full-Text PDF Breast cancer and hormone-replacement therapy: the Million Women StudyThe MWS collaborators1 provide data for oestrogen-progestagen HRT and oestrogen-only HRT, but not always for tibolone. Full-Text PDF Breast cancer and hormone-replacement therapy: the Million Women StudyThe accrual of more than a million women into a study1 is near unprecedented and a major accomplishment. However, the mere size of a study population does not imply that its results should be taken at face value, particularly when they do not concur with those of previous studies. The high risk estimates (RRs) for short-term current-users of, in particular, oestrogen-progestagen HRT and the lack of increased risk of breast cancer for women who stopped use less than 5 years ago are discrepant with those of other large studies,2–4 and we are concerned that biases inherent to the study design might play a part. Full-Text PDF Breast cancer and hormone-replacement therapy: the Million Women StudyFurther information is needed before the clinical significance of the results of the MWS1 can be properly assessed. Was breast cancer ruled out in each participant before HRT was prescribed? If so, were both physical examinations and mammography used? Until we know the interval between the last negative breast examination and the diagnosis of breast cancer, as well as the stage of breast cancer at the time of diagnosis, we cannot ascertain whether the detected cancers were incident or prevalent cases. Full-Text PDF
Referência(s)