A pharmacokinetic-pharmacodynamic study was performed in opioid-dependent patients in the Netherlands, who were currently treated with high doses of pharmaceutically prepared heroin on medical prescription. Besides intravenous heroin, heroin was prescribed for inhalation by "chasing the dragon" method. In this technique, heroin base is heated on aluminium foil, and heroin vapours are inhaled into the lungs. Not much is known about the pharmacokinetics profile and bioavailability of this specific administration method. Therefore, a study was performed on pharmacokinetics and pharmacodynamics of heroin inhalation and intravenous use. Eleven patients who injected heroin and 9 patients who inhaled heroin entered the study. They were on steady-state heroin treatment for at least 12 months. For safety reasons, there was no crossing-over between heroin injection or inhalation. In a double-blind randomised study, 67-100-150% of the regular heroin maintenance dose was administered to each patient. Maximal single heroin dose was 450 mg. Plasma concentrations of heroin and its metabolites 6-monoacetylmorphine, morphine and morphine-glucuronides were analysed using LC-MS-MS. Blood pressure, heart rate, skin temperature and reaction time were assessed. Furthermore, visual analogue scales regarding craving and appreciation of heroin effect were scored by the subjects. Both in inhaling and injecting patients, the areas under curve of heroin and all measured metabolites were linearly related to heroin dose. Mean C(max) of heroin and its metabolites were 2-6 times lower after inhalation, than after intravenous injection. Bioavailability (F) of heroin inhalation was estimated as 52% (95% CI 44-61%). Heroin was rapidly cleared from plasma. Cl/F was 930 l/hr (95% CI 799-1061 l/hr) after intravenous administration, and 1939 l/hr (95% CI 1661-2217 l/hr) after inhalation. Heroin Cl and Vd were correlated to body weight (R(2) 15-19%). Morphine-glucuronides levels were inversely related to creatinine clearance. After heroin administration, the reaction time was significantly prolonged with 28+/-5.3 msec. in injecting and 13+/-4.9 msec. in inhaling patients. Cardiovascular changes were only mild after heroin administration. Craving-scores declined immediately after heroin administration in both administration groups. Subjective heroin effect was rated more positively in heroin inhaling than in injecting patients, despite the lower C(max) levels following heroin inhalation. In both groups, in this blinded study heroin dose increments were more appreciated than dose reductions. Increments of 50% of the regular heroin dose did not cause any serious side effect.