The objective of this study was to compare cotinine concentrations in urine and saliva using gas chromatography (GC), high-performance liquid chromatography (HPLC) and enzyme-linked immunosorbent assay (ELISA). Ninety-four subjects were selected (27 smokers and 67 non-smokers) and interviewed using questionnaire. Of the non-smokers, 39 had been exposed to environmental tobacco smoke (ETS) and 28 had not been exposed to ETS. Cotinine levels among smokers were highest using all three measurements, followed by ETS exposed subjects and non-smokers. Cotinine levels in urine, using HPLC, correlated significantly with levels measured using ELISA (r=0.92) and GC-nitrogen-phosphorus detection (NPD) (r=0.92). Salivary cotinine levels measured using ELISA did not correlate significantly with either HPLC (r=0.37) or GC-NPD (r=0.33) measurements. Multiple regression models were used to adjust for age, gender, drug use and health status, and it was found that cotinine levels in urine and saliva were significantly correlated with smoking pack-year. The authors conclude that urinary cotinine concentration is a more accurate biomarker for ETS than salivary cotinine concentration.