Testosterone (T) in a hydroalcoholic gel has been developed as an effective and convenient open system for transdermal delivery of the hormone to men. Because the gel can be applied either to small or large areas of skin, it was important to assess whether the skin surface area on which the gel was applied was an important determinant of serum T levels. To answer this question, the pharmacokinetics of a transdermal 1% hydroalcoholic gel preparation of T was studied in nine hypogonadal men. The subjects applied in random order a 25-mg metered dose of T gel either four times at one site (left arm/shoulder) or at four different sites (left and right arms/shoulders and left and right abdomen) once daily (6-8 min) for 7 consecutive days. After 7 days of washout, each subject was then crossed over to the opposite regimen for another 7 days of treatment. Serum samples were collected for measurements of T, 5alpha dihydrotestosterone (DHT), and estradiol before, during (days 1, 2, 3, 5, and 7), and after (days 8, 9, 11, 13, and 15) application of T gel. Multiple blood samples were drawn on the 1st and 7th day after gel application; single samples were obtained just before the next T gel application on other days (24 h after the previous gel application). The T gel dried in less than 5 min, left no residue, and produced no skin irritation in any of the subjects. Mean serum T levels, irrespective of application at one site or four sites followed the same pattern: rising to 2- to 3- and 4- to 5-fold above baseline at 0.5 and 24 h after first application, respectively. Thereafter, serum T levels reached steady state and remained at 4- to 5-fold above baseline (at the upper limit of the normal adult range) for the duration of gel application and returned to baseline within 4 days after stopping application. The application of T gel at four sites (application skin area approximately four times that of one site) resulted in a mean area under the curve (AUC0-24h) for serum T levels on the 7th day (868 +/- 72 nmol*h/L, mean +/- SEM), which was 23% higher but not significantly different (P = 0.06) than repeated application at one site (706 +/- 59 nmol*h/L). This could be due to the limited number of subjects studied (n = 9). Mean serum DHT levels followed the same pattern as serum T, achieving steady-state levels by 2 days. The mean concentration of serum DHT on the 7th day was significantly higher after application at four sites (9.15 +/- 1.26 nmol/L, P < 0.05) than at one site (6.9 +/- 0.77 nmol/L). These serum DHT levels were at or above the normal adult male range. Serum DHT:T ratio was not significantly altered by T gel application. Serum estradiol levels followed the same pattern as serum T and showed no significant difference between the one- or four-site application. We conclude that transdermal daily application of 100 mg T gel resulted in similar steady levels of serum T. The surface area of the skin to which the gel was applied had only a modest impact on serum T and DHT levels. Mean serum levels of T and DHT was higher by 23% and 33%, respectively, despite application of the gel to four times the skin area in the four sites compared with the one site group. Because of the greater dosage flexibility provided, hydroalcoholic T gel application over multiple sites seems to be an effective and nonskin-irritating method of transdermal T delivery for hypogonadal men. Dose-ranging studies are required to determine dosage regimens for T gel application as a replacement therapy in hypogonadal men.