Recent epidemiological evidence relevant to the clinical management of the menopause.

BACKGROUND: The 2003 Workshop of the International Menopause Society considered the epidemiological evidence collected up to that time on the effects of female hormone therapy (HT). New evidence relevant to the clinical management of the menopause has since been published. OBJECTIVES: To summarize the new evidence, to offer critiques of important recently published studies, and to consider the implications for clinical practice. CARDIOVASCULAR DISEASE: Recent evidence from two studies, the Women's Health Initiative (WHI) clinical trial, and an observational component of the WHI, suggests that combined hormone therapy (estrogen plus progestin) (CHT) initially increases the risk of coronary heart disease (CHD), stroke, and venous thromboembolism (VTE), followed by a decline. For CHD, the hazard ratio exceeds 1.0 during the first year of follow-up, followed by a progressive decline to <1.0 after >5 years. Other studies show the same trend. BREAST CANCER: In the WHI data, recent evidence suggests that estrogen therapy (ET) reduces the overall risk of breast cancer, predominantly ductal and localized cancer. Evidence from the Million Women Study (MWS) now suggests that the previously reported association of HT with breast cancer is concentrated on tumors with lobular or tubular histology; the risk of ductal cancer is also increased, but to a lesser degree. The risks of these outcomes are higher for CHT than for ET. Other recent studies broadly accord with the MWS observations. OTHER OUTCOMES: Among CHT recipients, the WHI findings of reduced risks of fractures and colorectal cancer, and an increased risk of VTE, remain unchanged. Evidence from other studies now suggests that protracted exposure to CHT may increase the risk of ovarian cancer, and reduce the risk of endometrial cancer. INTERPRETATION: The recently published WHI findings for CHD and breast cancer are of major importance. For CHD, detection bias may have resulted in systematic overestimation of the duration-dependent hazard ratios. If so, there may be no initial increase in the risk, and prolonged use may be associated with a decreased risk. The hypothesized protective effect of HT may have been missed in the WHI study. For breast cancer, the WHI evidence now suggests a protective effect of ET. Tumors with lobular or tubular histology tend to be small, slow-growing, low-grade, and well differentiated. Such tumors may be more susceptible to detection bias, and that bias has not been ruled out as an alternative explanation of the higher risks among CHT recipients, observed in the MWS. The possibility of detection bias in that study, and in other observational studies, is supported by the decreased risk of breast cancer observed among ET recipients in the WHI clinical trial. Based on the present evidence, it is impossible to determine whether HT, or specific forms of HT, increase, decrease, or have no effect on the overall risk of breast cancer, or of specific types of breast cancer. Other evidence raises the possibility that prolonged CHT may increase the risk of ovarian cancer, and decrease the risk of endometrial cancer. Additional studies are needed to confirm those findings. If, as now seems possible, CHT in fact reduces the risk of CHD, and has little or no effect on the risk of breast cancer, or if ET decreases the risk, the clinical and public health implications would be major. However, the picture is confused. In view of new, but uncertain, findings concerning CHD and breast cancer, clinicians will have to continue to use clinical judgment, informed by a critical evaluation of the epidemiological evidence, in the management of the menopause.