WHEN RECRUITMENT FOR THE WOMEN’S HEALTH Initiative (WHI) began in 1993, hormone therapy (HT) was prescribed for a variety of reasons that ranged from the management of menopausal symptoms to the prevention of chronic disease, and the WHI focused on documenting the risks and benefits of HT use for chronic disease prevention. In 2004, one of the first publications from the WHI reported the balance of risks and benefits during active intervention with unopposed estrogen among women with previous hysterectomy. Based on data collected through the end of the trial’s intervention phase, women randomly assigned to estrogen had a significantly increased risk of stroke and reduced risk of hip fracture and possibly breast cancer compared with women receiving placebo. There was no overall effect of estrogen on a global index of risks and benefits including coronary heart disease. These findings were revolutionary and changed practice. Today, indications for HT are narrow, and many physicians take pause before performing elective bilateral salpingooophorectomy during hysterectomy. Short-term unopposed estrogen therapy is still a mainstay in managing menopausal symptoms among women with previous hysterectomy, although questions remain regarding the safety of this treatment, including whether there is a safe duration of estrogen use and whether there are long-term risks and benefits of estrogen therapy that persist after cessation. Once again, data from the WHI study in this issue of JAMA help to provide guidance. To assess the long-term effects of unopposed estrogen therapy and the balance of risks and benefits over time, LaCroix and colleagues report follow-up of women who completed the intervention phase of the WHI estrogen-alone study. Seventy-eight percent of eligible women gave consent for continued observational follow-up through a mean of 10.7 years from baseline. During follow-up, the increased risk of stroke observed previously did not persist after cessation of estrogen therapy, whereas the previously observed significant reduction in hip fractures was eliminated. Other risks and benefits associated with estrogen therapy during the intervention phase were not maintained; however, the reduced incidence of breast cancer persisted. This finding is inconsistent with a longstanding, corroborated body of evidence and raises the possibility that other important factors modify documented risks and benefits of estrogen therapy among these long-term WHI participants. Building on previous evidence that highlights the importance of age at HT initiation, LaCroix and colleagues postulate that an age effect may also underlie some of their findings. For example, the authors expected fewer adverse effects among women who initiated estrogen at the time of menopause, and they found that for every 10 000 women aged 50 to 59 years taking estrogen, there were 12 fewer myocardial infarctions, 13 fewer deaths, and 18 fewer adverse events compared with those taking placebo, and they conclude the overall benefits of estrogen therapy may be greater among younger women. In contrast, a recent report by Beral et al from the Million Women Study demonstrates an adverse effect of postmenopausal estrogen use on breast cancer risk, with a significantly increased risk among women beginning therapy within 5 years of menopause and little or no increased risk among those beginning therapy 5 or more years after menopause. These findings derive from an analysis of 15 759 incident breast cancer cases diagnosed during 4 million patient-years of follow-up. Sixty-eight percent of women enrolled in the WHI were older than age 60 years at randomization. Given this fact and the findings from the Million Women Study, an important question that emerges is whether the WHI population is appropriate for reaching definitive conclusions regarding younger women and the risk of breast cancer associated with HT. In addition to a potential age effect on the risk-to-benefit profile of HT, overall duration of HT use remains a major concern. The median adherent time (defined as women taking 80% of study pills) among women in the WHI estrogenalone group was 3.5 years. Thus, the WHI results do not address the balance of risks and benefits associated with longerterm estrogen use. Longer unopposed estrogen use may