Although evidence seems to indicate favorable effects of hormone replacement therapy (HRT) on cognitive functions and mood in elderly healthy and demented women, some questions remain. For instance, the nature of the long term effect of HRT, e.g. in preventing cognitive decline is still unclear. In this respect the addition of progestagens in combined HRT has been mentioned to oppose some of the beneficial effects of estrogens. The present paper aims to illuminate these questions and presents two studies. Tn the first study, the long term effects of combined HRT in healthy postmenopausal women was investigated using a parallel groups (HRT-users vs. controls) design. HRT subjects were always tested during the estrogen-progestagen phase. Results indicated that after 6 and 12 months, women in the HRT-treatment group had higher scores on several indicators of the subjective feeling of well being (sleep, physical and psychological complaints) than matched controls, although at baseline both groups were not severely impaired. Effects of HRT on memory functions were seen when HRT treated subjects were compared with their own baseline functioning, but not when compared with controls. Hence, the addition of progestagen did not oppose the effects of estrogens on subjective feelings of well being or on memory. Our second (case-control) study involved women of middle-age who were unaware of the purpose of the experiment. mo positive effects of HRT use on subjective scales of well being or on memory were found. However, women with HRT were faster on basic sensorimotor speed tasks as compared with controls. It should be kept in mind that double blind testing in an experimental study is difficult due to withdrawal bleeding and the reduction of flushes. Expectancy effects may hav confounded the results of the first study. However, our findings indicate that the use of a particular design and type of memory test can explain the controversial results of studies into the effect of HRT on cognitive function. Furthermore, it was concluded that HRT has a global activating, instead of specific direct effect on cognitive functions. (C) 1998 Elsevier Science Ltd. All rights reserved.