We examined the effect of smoking on the treatment response in serum estrogens, serum lipids and lipoproteins, and bone mineral content in 110 postmenopausal women treated for 2 years with either percutaneous or orally administered combined hormones or placebo and followed up by examinations every 3 months. Serum estradiol and estrone levels during oral hormone administration were lower in smokers than in nonsmokers, whereas no differences related to smoking habits were observed during percutaneous hormone administration. Serum total and low-density lipoprotein cholesterol were significantly reduced in both smokers and nonsmokers receiving hormones, but the response in smokers was only half that observed in nonsmokers (not significant). When the impact of the route of hormone administration was examined in relation to smoking habits, the response in serum total and low-density lipoprotein cholesterol in smokers receiving oral hormones was significantly lower (p less than 0.05) than that observed in nonsmokers. No differences in serum lipids or lipoproteins were observed between smokers and nonsmokers receiving percutaneous hormones. The response in bone mineral content in smokers and nonsmokers receiving percutaneous hormones and placebo was not significantly different, although the overall response differed significantly in the two groups (p less than 0.001). In contrast, the response in smokers receiving oral hormones was significantly lower than that observed in the corresponding nonsmokers (p less than 0.01). We conclude that smoking greatly affects the response on circulating levels of estrogens in postmenopausal women treated with orally administered hormone replacement therapy and that the subsequent treatment response on serum lipids and lipoproteins and on bone mineral content is reduced accordingly. The study suggests that alternative routes of administration should be considered when postmenopausal estrogen therapy is instituted in women who smoke.