Risk-benefit relation differs between populations and individuals

Papers p 1858

Hormone replacement therapy is increasingly advocated not just for short term treatment of menopausal symptoms but as long term prophylactic therapy against heart disease, osteoporosis, even Alzheimer's disease—indeed, as the solution to many of the problems of ageing women.1 Should universal hormone replacement therapy be recommended in asymptomatic healthy postmenopausal women?

Many clinicians now take it as established that postmenopausal hormone therapy protects against coronary heart disease in women. However, this is not based on data from randomised trials with coronary end points. Hemminki and McPherson attempted to see whether useful information on the incidence of cardiovascular diseases and cancer could be obtained from published clinical trials which studied other outcomes of postmenopausal hormone therapy.2 Despite pooling data from all the small trials they found no convincing evidence one way or the other. Most striking is the tiny size and short duration of most of the trials, which clearly had inadequate power to examine clinical events. Even the largest, the Postmenopausal Estrogens and Progestins Intervention (PEPI) trial,3 had fewer than 200 women in each treatment arm. Apart from the Nactigall study,4 the median duration was not much more than one year. There was substantial uncertainty about definition and ascertainment of events, which were often reported as dropouts or asides and not as the primary focus of the trials. Given the variable quality of these trials and procedures, the small difference in numbers of events could easily have arisen from biases in ascertainment and reporting. Why then, are we so convinced of the cardioprotective effect of oestrogens? …