The psychostimulant drug of abuse, cocaine (benzoylecgonine methyl ester), is rapidly metabolized by cleavage of its two ester groups, to give benzoylecgonine (BE) and ecgonine methyl ester, and by N-demethylation, to give N-norcocaine (NC). The recent use of [N-methyl-11CH3]cocaine to image brain cocaine binding sites with positron emission tomography (PET) raises the question of whether PET images partially reflect the distribution and kinetics of labeled cocaine metabolites. We prepared [O-methyl-11CH3]cocaine by methylation of the sodium salt of BE with [11C]CH3I, and showed that PET baboon brain scans, as well as regional brain kinetics and plasma time-activity curves corrected for the presence of labeled metabolites, are nearly identical to those seen with [N-methyl-11CH3]cocaine. This strongly suggests that 11C metabolites do not significantly affect PET images, because the metabolite pattern is different for the two labeled forms of cocaine. In particular, nearly half the 11C in blood plasma at 30 min was [11C]CO2 when [N-methyl-11CH3]cocaine was administered, whereas [11C]CO2 was not formed from [O-methyl-11CH3]cocaine. Only a trace of [11C]NC was detected in plasma after [O-methyl-11CH3]cocaine administration. Nearly identical brain PET data were also obtained when 4'-[N-methyl-11CH3]fluorococaine and 4'-[18F]fluorococaine (prepared by nucleophilic aromatic substitution from [18F]fluoride- and 4'-nitrococaine) were compared with [N-methyl-11CH3]cocaine. In vitro assays with rat brain membranes showed that cocaine and 4'-fluorococaine were equipotent at the dopamine reuptake site, but that 4'-fluorococaine was about 100 times more potent at the 5-hydroxytryptamine reuptake site.(ABSTRACT TRUNCATED AT 250 WORDS)