(1) The WHI study was published in 2002: a randomised double-blind placebo-controlled clinical trial in more than 16 000 women with an average age of 63 years at enrollment. The paper reports data on the
long-term adverse effects of combined equine
estrogen-
progestin hormone replacement therapy, taken for 5 years. (2) On average, a yearly excess of 19 severe adverse events per 10 000 women occurred in the
estrogen-
progestin group. Relative to the placebo group, there were an extra 8
pulmonary embolisms, 7 coronary events, 8
strokes and 8 cases of invasive
breast cancer. In contrast, there were 6 fewer
colorectal cancers and 5 fewer
hip fractures in the active treatment group. (3) The differences in the frequency of coronary events and
venous thromboembolism emerged after the first year of treatment, while the curves for
stroke and
breast cancer diverged after the second and fifth years, respectively. (4) The overall mortality rate did not differ between the two groups. (5) A placebo-controlled trial of the same
hormone combination (HERS trial), given for 4.1 years as secondary prophylaxis against
coronary heart disease was published in 1998. The
drug was ineffective during the trial, and during unblinded post-trial follow-up of 2 321 women for an average of 2.7 years (HERS II study). (6) The
estrogen-
progestin combination used in these trials did not reduce the risk of
coronary heart disease (in primary or secondary prophylaxis) or the risk of
stroke. On the contrary, both risks increased. (7) The increased incidence of
deep venous thrombosis and/or
pulmonary embolism associated with
estrogen-progestin replacement therapy was confirmed in these trials, even among women with no relevant history. (8) The WHI trial also confirmed the increased risk of
breast cancer in women on
hormone replacement therapy, but did not study its impact on outcome or mortality. (9) The WHI trial confirmed the beneficial impact of
estrogen-progestin combination therapy on the risk of
osteoporotic fracture. An average of 5
hip fractures were avoided each year per 10 000 women treated (10 versus 15 observed cases per 10 000 women per year), together with 6 symptomatic vertebral fractures (9 versus 15 cases) and 44
osteoporotic fractures (147 versus 191 cases). It is not known whether the benefit observed at the end of the trial persisted after the end of treatment. (10) In practice, the decision to prescribe
hormone replacement therapy, and the optimal
duration of treatment, must be weighed up according to each individual's risk factors. And the decision to treat or not to treat must be regularly re-assessed. Women must be informed of the potential risks and benefits, and must be monitored. They should also be advised not to use less well assessed treatments such as
phytoestrogens,
DHEA and
tibolone. (11) Health authorities, especially in Europe, must organise comparative trials to assess the benefits and risks of other
hormone combinations used by perimenopausal and postmenopausal women.