Adolescent tobacco smokers have a higher prevalence of marijuana (MJ) smoking than adolescents who do not smoke tobacco. As part of an adolescent smoking cessation trial, we examined whether MJ smoking, and specifically "blunt" (gutted cigars filled with MJ) smoking, elevated participants' likelihood of a false indication of cigarette smoking on the basis of breath carbon monoxide (CO) testing. Using clinical data from 37 adolescents (mean age 15.1+/-1.4 years, 78% female) who participated in a smoking-cessation trial in Baltimore between 1999 and 2002, and who on at least one occasion, reported abstinence from tobacco smoking for at least 7 days, we analyzed 146 cigarette-abstinent-visit exhaled CO concentrations classified into blunt occasions (12 participants, 33 visits), nonblunt MJ occasions (seven participants, 20 visits), and non-MJ occasions (27 paricipants, 93 visits). Repeated-measures logistic regression revealed that blunt occasions were associated with CO > or = 8 ppm, compared to nonblunt occasions (p = 0.013). Blunt occasions also tended to be associated with the more youth-appropriate cutoff CO > or = 6 ppm, compared to non-MJ occasions (p =0.054). Blunt smoking impacted the interpretation of measures of exhaled CO for tobacco cessation.