We compared the transdermal and subdermal routes of estrogen administration with respect to the constancy of estrogen delivery and metabolic effects. Twenty postmenopausal women were randomized to receive either two 25 mg estradiol pellets subdermally (n = 10) or a 0.1 mg estradiol transdermal patch twice weekly (n = 10). Blood was sampled at 0, 2, 4, 6, 8, 12, 24, and 72 hours and 1, 2, 4, 8, 12, 16, 20, and 24 weeks (fasting samples at 0, 12, and 24 weeks), and a fasting urine was obtained after diuresis at 0, 12, and 24 weeks. In a 72-hour profile, serum estradiol levels (mean +/- SE) were highest at 24 hours (179 +/- 20 pg/ml) and fell to 139 +/- 16 pg/ml at 72 hours in the pellet group. In the patch group, estradiol levels rose rapidly to 152 +/- 33 pg/ml at 4 hours, remained relatively constant over 8 hours, and fell to 46 +/- 10 pg/ml at 72 hours. At 1 week, estradiol levels in the pellet group were 113 +/- 12 pg/ml and remained relatively constant for 24 weeks. In contrast, estradiol levels in the patch group were 52 +/- 11 pg/ml at 1 week and then varied widely until 24 weeks, when the levels were 89 +/- 26 pg/ml. The mean estradiol/estrone ratio ranged between 1 and 2.5 in both groups but fluctuated widely in the patch group. Follicle-stimulating hormone was suppressed in both groups; however, the decrement in the pellet group was greater (p less than 0.002). There was a significant increase in high-density lipoprotein cholesterol and a decrease in total cholesterol/high-density lipoprotein cholesterol at 12 weeks with the pellet but only at 24 weeks with the patch. The urinary calcium/creatinine ratio was reduced more consistently with the pellet than with the patch. Hot flushes were eliminated in all subjects.