OBJECTIVE There are advantages and disadvantages with all of the presently available types of testosterone replacement for hypogonadal men. We performed this investigation to establish detailed data about the pharmacokinetics, pharmacodynamics, feasibility and side-effects of subcutaneously implanted testosterone (T) pellets. METHODS In a single-dose, open-label, non-randomized study, 6 T-pellets, each containing 200 mg of fused crystalline T, were implanted in the subdermal fat tissue of the lower abdominal wall of 14 hypogonadal men. Blood samples for determination of T, LH, FSH, 5 alpha-dihydrotestosterone (DHT), sex hormone binding globulin (SHBG) and oestradiol (E2) were obtained at 0, 0.5, 1, 2, 4, 8, 12, 24, 36, 48 hours and on day 21 after implantation and then every 3 weeks until day 189, and on days 246 and 300 during follow-up. In another 36 hypogonadal men the feasibility and side-effects of T-pellets were evaluated. METHODS Fourteen patients participated in the detailed pharmacokinetic study and another 36 patients in the assessment of feasibility and side-effects. All patients (age range 18-61 years) suffered from primary or secondary hypogonadism (T < 3.6 nmol/l). RESULTS The pharmacokinetic study in 14 hypogonadal men revealed an initial short-lived burst release of T with a peak concentration of 49.0 +/- 3.7 nmol/l at 0.5 +/- 0.13 days which was followed by a stable plateau lasting until day 63 (day 2, 35.2 +/- 2.3; day 63, 34.8 +/- 2.6 nmol/l). Thereafter serum T gradually declined and was close to baseline concentrations on day 300. Apparent terminal elimination half-life (t1/2) was 70.8 +/- 10.7 days and apparent mean residence time 87.0 +/- 4.5 days. On average, serum T was below 10 nmol/l after 180 days. Absorption of T followed a zero-order release kinetic with an absorption half-time of 74.7 days (95% confidence interval: 71.1-78.5) and was almost complete by day 189 (95.9 +/- 0.84%). Serum DHT and E2 were significantly elevated from day 21 to day 105 and correlated significantly with T (DHT, r = 0.65, P < 0.0001, E2, r = 0.67, P < 0.0001). SHBG was significantly decreased from day 21 to day 168. In 6 men with primary hypogonadism T suppressed LH and FSH to the eugonadal range from day 21 to 126 and 42 to 105, respectively, with nadirs occurring at day 84 (LH) and day 63 (FSH). LH and FSH were highly inversely correlated with T (r = -0.47 and -0.57). The only side-effect observed during 112 implantations in the total group of 50 men were 6 local infections (5.4%) leading to extrusion of 5 pellets in 3 men. When given the choice, all patients except one preferred T-pellets to their previous T medication for permanent substitution therapy. CONCLUSIONS T-pellets are the androgen formulation with the longest biological action and strongest pharmacodynamic efficacy in terms of gonadotrophin suppression. The pharmacokinetic features are advantageous compared to other T preparations and the patient acceptance is high.