Due to the technological difficulties associated with delivering reproducible doses by the smoking route, there have been few clinical studies characterizing the pharmacokinetic and pharmacodynamic profiles of smoked drugs of abuse. This research was designed to represent an integrated pharmacokinetic and pharmacodynamic approach for the delivery of smoked heroin and cocaine to human subjects residing in a controlled clinical research unit. Smoked drugs were administered utilizing a computer assisted smoking device which delivered drug in a single puff with minimal pyrolysis and no loss due to main or sidestream smoke. Volunteers were administered four smoked doses and four intravenous doses of heroin or cocaine in an ascending dose design with random placement of placebo. Physiological, subjective and performance measures were collected concurrently with biological samples. Blood, plasma and saliva were analyzed by solid phase extraction-gas chromatography/mass spectrometry. Heroin appeared rapidly in blood after smoking, peaked at 1-5 min and declined to the limit of detection (LOD) of the assay by 30 min. Mean elimination half-lives for heroin, 6-acetylmorphine and morphine were 3.3 min, 5.4 min, and 18.8 min, respectively, after smoking. Physiological measures, such as pupil diameter, demonstrated a counterclockwise hysteresis compared with heroin levels. This work demonstrated that smoked heroin produces similar pharmacological effects as are produced by the intravenous administration of heroin. A similar pharmacokinetic and pharmacodynamic study was performed with human subjects smoking cocaine. Physiological, subjective and performance measures of drug effect were correlated with cocaine plasma concentrations after smoked cocaine. Some physiological effects (pupil diameter, heart rate) demonstrated a delay between peak plasma concentrations and maximal response whereas other measures of drug effect demonstrated no delay (systolic and diastolic blood pressure, "feel", "liking" and "good" drug subjective effects). Several subjective measures of drug effect declined to baseline levels at a faster rate than decreasing plasma cocaine concentrations, demonstrating acute tolerance to the effects of cocaine. The research design also included testing of saliva as an alternate biological matrix for the determination of pharmacokinetic parameters of heroin and cocaine. Saliva drug concentrations were compared to blood concentrations. After drug administration, peak heroin concentrations were achieved rapidly in saliva, between 2-5 min after intravenous administration and at 2 min after smoked heroin. Saliva concentrations of heroin declined rapidly after intravenous administration, reaching the LOD by 60 min, compared with detection times of 4-24 h after smoking. Cocaine was the major analyte detected in saliva after smoking and intravenous administration. The thermal degradation product of cocaine, anhydroecgonine methyl ester, was detected in saliva after smoking with peak concentrations occurring at 2 min. (Abstract shortened by UMI.)