Abstract Although expired carbon monoxide (CO) and plasma cotinine (COT) have been validated as biomarkers of self-reported cigarettes per day (CPD) in heavy smoking Caucasians, their utility in light smokers is unknown. Further, variability in CYP2A6, the enzyme that mediates formation of COT from nicotine and its metabolism to trans-3′-hydroxycotinine (3HC), may limit the usefulness of COT. We assessed whether CO and COT are correlated with CPD in African-American light smokers (≤10 CPD, n = 700), a population with known reduced CYP2A6 activity and slow COT metabolism. We also examined whether gender, age, body mass index, smoking mentholated cigarettes, or rate of CYP2A6 activity, by genotype and phenotype measures (3HC/COT), influence these relationships. At baseline, many participants (42%) exhaled CO of ≤10 ppm, the traditional cutoff for smoking, whereas few (3.1%) had COT below the cutoff of ≤14 ng/mL; thus, COT seems to be a better biomarker of smoking status in this population. CPD was weakly correlated with CO and COT (r = 0.32-0.39, P < 0.001), and those reporting fewer CPD had higher CO/cigarette and COT/cigarette, although the correlations coefficients between these variables were also weak (r = −0.33 and −0.08, P < 0.05). The correlation between CPD and CO was not greatly increased when analyzed by CYP2A6 activity, smoking mentholated cigarettes, or age, although it appeared stronger in females (r = 0.38 versus 0.21, P < 0.05) and obese individuals (r = 0.38 versus 0.24, P < 0.05). Together, these results suggest that CO and COT are weakly associated with self-reported cigarette consumption in African-American light smokers, and that these relationships are not substantially improved when variables previously reported to influence these biomarkers are considered. (Cancer Epidemiol Biomarkers Prev 2009;18(12):3426–34)