Health Risks and Benefits 3 Years After Stopping Randomized Treatment With Estrogen and Progestin

JAMA. 2008 Mar 5;299(9):1036-1045, G Heiss, R Wallace, GL Anderson, A Aragaki, SAA Beresford, R Brzyski, RT Chlebowski, M Gass, A LaCroix, JE Manson, RL Prentice, J Rossouw, ML Stefanick, for the WHI Investigators

The Women’s Health Initiative (WHI) study, evaluated the combined effects of conjugated equine estrogen (CEE) and medroxyprogesterone acetate (MPA) in 16,608 postmenopausal, predominantly healthy, women aged 50 to 79 years. The trial was halted after 5.6 years when planned interim analysis results showed that, compared with placebo, HRT was associated with significantly increased incidences of invasive breast cancer, cardiovascular disease (CVD), stroke, venous thromboembolism (VTE), and pulmonary embolism (PE). However, women taking CEE/MPA had lower risks of bone fracture and colorectal cancer. Here, the results of a planned 8.5-year follow-up analysis, including 15,730, women are reported. This postintervention phase of the WHI incorporated data collected during a mean 2.4 year period commencing after the trial was stopped on July 7, 2002, until the planned termination date of March 31, 2005. 

Results showed that the increased risks of CVD, VTE, and PE associated with CEE/MPA use completely dissipated during this period. Conversely, the incidence rate of “all cancers” increased by 1.56% in CEE/MPA treated-women compared with 1.26% for placebo-treated women from the time of the intervention to the postintervention analysis (hazard ratio [HR] = 1.24, 95% confidence interval [CI], 1.04-1.48). Colorectal cancer rates were equivalent in both treatment groups; therefore, the benefit originally observed with CEE/MPA treatment was lost. For endometrial cancer, there was an insignificantly lower risk in CEE/MPA- vs placebo-treated women (HR = 0.75, 95% CI, 0.40-1.43). The difference in the incidence of fractures between the 2 groups was almost insignificant. Postintervention analysis of all-cause mortality showed only a nominally increased risk in CEE/MPA- vs placebo-treated women (HR = 1.15, 95% CI, 0.95-1.39).

Compared with the intervention analysis, the HR for malignancy in CEE/MPA-treated women increased during the postintervention analysis, rising from1.03 (95% CI, 0.92-1.15) to 1.24 (95% CI, 1.04-1.48, P = .08). The majority of deaths among the study participants were cancer-related and these were higher in the CEE/MPA- than the placebo-treated women (101 vs 69, respectively). Only 27 vs 16 deaths, respectively, were attributed to the prespecified cancer outcomes (breast, colorectal, endometrial, or ovarian cancer). Among those classified as “other cancers”, most were lung cancers which occurred in 33 and 15 of the CEE/MPA- and placebo-treated women, respectively.

In conclusion, new patterns of health risks and benefits emerged during the postintervention analysis of the WHI study. Within 3 years since the intervention analysis, the hip fracture rates evened out between the CEE/MPA- and placebo-treated women while the risk of cancer increased slightly in the CEE/MPA-treated women. Overall, these results suggest that the benefits of postmenopausal women receiving CEE/MPA treatment do not outweigh the risks, based on an adverse trend in all-cause mortality associated with CEE/MPA use.