The clinical consequences of androgen deficiency in human immunodeficiency virus (HIV)-infected women remain underappreciated. The pharmacokinetics of transdermally administered testosterone in premenopausal women and HIV-infected women have not been studied. In this study we compared the pharmacokinetics of a novel testosterone matrix transdermal system (TMTDS) in healthy premenopausal women and women infected with HIV. Eight menstruating HIV-infected women, 18-50 yr of age, who had been receiving stable antiretroviral therapy, including a protease inhibitor, for at least 12 weeks and nine healthy, menstruating women of comparable age were enrolled. After baseline sampling during a 24-h control period in the early follicular phase (days 1-6), two TMTDS patches were applied with an expected delivery rate of 300 microg testosterone daily over an application period of 3-4 days. After 72 h, the patches were removed, a second set of two patches was applied, and blood samples were drawn over 96 h. Baseline serum total and free testosterone levels were lower in HIV-infected women than in healthy women. A diurnal rhythm of testosterone secretion, with higher levels in the morning and lower levels in the late afternoon, was apparent in both groups of women. Free testosterone levels were in the midnormal range at baseline in healthy women and increased above the upper limit of normal during TMTDS application. In HIV-infected women, free testosterone levels were in the low normal range at baseline and rose into the upper normal range during patch application. Serum total testosterone levels increased into the midnormal range in HIV-infected women and into the upper normal range in healthy women during patch application. The mean increments in free and total testosterone levels were significantly lower in HIV-infected women than in healthy women. Testosterone bioavailability, expressed as the mean +/- SEM baseline-subtracted area under the total testosterone curve, was significantly greater in healthy women than in HIV-infected women [3323 +/- 566 ng/dL x h (115 +/- 20 nmol/L x h) vs. 1506 +/- 316 ng/dL x h (52 +/- 11 nmol/ L x h); P = 0.016]. Assuming a daily testosterone delivery rate of 300 microg/day, the apparent plasma clearance was significantly higher in HIV-infected women than in healthy women (2531 +/- 469 vs. 1127 +/- 217 L/day1 P = 0.022), respectively. There was no significant change from baseline in serum LH, sex hormone-binding globulin, and estradiol levels in either group. Serum FSH levels showed a greater decrease from baseline in healthy women. A regimen of two testosterone patches applied twice a week can maintain serum total and free testosterone levels in the mid- to upper normal range, respectively, in HIV-infected women with low testosterone levels. During TMTDS application, the increments in serum total and free testosterone levels are lower in HIV-infected women than in healthy women, presumably due to increased plasma clearance or decreased absorption. Further studies are needed to assess the effects of physiological androgen replacement in HIV-infected women.