Estriol is regarded as a mild and brief-acting hormone and, as such, has been prescribed for over four decades as a therapeutic agent for postmenopausal women with urogenital disorders. The dosages administered and the results obtained have varied considerably in earlier studies. Pharmacokinetic studies of estriol and the effects induced on the target tissues have been contradictory. The pharmacokinetics of unconjugated estriol after vaginal and oral medication was studied. One mg estriol applied vaginally resulted in plasma estriol concentrations similar to those obtained when 10 mg estriol was given orally. The clinical importance of the enterohepatic recirculation on plasma estriol concentrations after orally administered estriol was demonstrated. Factors that influence the pharmacokinetics of estriol are the time of estriol administration, fat-rich food, as well as the spacing of meals. Thus, the enterohepatic recirculation of estriol affects the plasma estriol elevation time and thereby the potency of estriol. The enterohepatic recirculation seemed to be less important in vaginal administration of estriol as the first liver passage is circumvented. Instantaneous and substantial conjugation of estriol in the liver is thereby avoided. By cytological assessment of the vaginal and the urethral epithelium in 40 patients treated in a geriatric hospital, the commonly recommended maintenance dosages of estriol (0.5 mg vaginally 1-2 times per week) were found to be too low to cure vaginal and urethral atrophy in very old women. This study suggests that estriol is a biologically potent estrogen by both oral and vaginal administration. The route and the time of administration, the food habits as well as the dosages given are factors which contribute to the estrogenicity of estriol. In clinical use, vaginal treatment seems preferable because the first liver passage is avoided. Vaginal administration allows a more standardized estriol regimen, while the oral treatment is more difficult to tailor for suit most women's needs.