BACKGROUND: Cannabis use results in impaired driving and an increased risk of motor vehicle crashes. Cannabinoid concentrations in blood and other matrices can remain high long after use, prohibiting the differentiation between acute and chronic exposure. Exhaled breath has been proposed as an alternative matrix in which concentrations may more closely correspond to the window of impairment; however, efficient capture and analytically sensitive detection methods are required for measurement. METHODS: Timed blood and breath samples were collected from 20 volunteers before and after controlled administration of smoked cannabis. Cannabinoid concentrations were measured using LC-MS/MS to determine release kinetics and correlation between the 2 matrices. RESULTS: DELTA.sup.9-Tetrahydrocannabinol (THC) was detected in exhaled breath for all individuals at baseline through 3 h after cannabis use. THC concentrations in breath were highest at the 15-min timepoint (median = 17.8 pg/L) and declined to <5% of this concentration in all participants 3 h after smoking. The decay curve kinetics observed for blood and breath were highly correlated within individuals and across the population. CONCLUSIONS: THC can be reliably detected throughout the presumed 3-h impairment window following controlled administration of smoked cannabis. The findings support breath THC concentrations as representing a physiological process and are correlated to blood concentrations, albeit with a shorter window of detection.