Elimination of interfering substances in urine by solid phase extraction (SPE) prior to analysis resulted in 10-fold improvement in the sensitivity of atrazine mercapturate (AM) enzyme-linked immunosorbent assay (ELISA) compared to previous reports. Of the two tested SPE systems, Oasis HLB and MCX, the mixed-mode MCX gave good recoveries (82%) of AM in spiked samples measured by ELISA, whereas the reverse-phase HLB phase was not compatible with the immunochemical method. At relatively high concentrations of urinary AM (>20 ng mL(-1)), sample dilution was effective enough for the elimination of interfering substances. The new liquid chromatography-mass spectrometry (LC-MS) method developed for AM utilizes online-SPE with Oasis HLB, column switching and a stable-isotope internal standard. The limit of quantification (0.05 ng mL(-1)) indicates improved sensitivity compared with most previously published LC-MS methods for AM. Validation of all three methods, LC-MS, ELISA+SPE and ELISA+dilution with spiked urine samples showed good correlation between the known and measured concentrations with R2 values of 0.996, 0.957 and 0.961, respectively. When a set (n=70 plus 12 blind duplicates) of urine samples from farmers exposed to atrazine was analyzed, there was a good agreement (R2=0.917) between the log normalized data obtained by ELISA+SPE and LC-MS. High correlation among the data obtained by the two tested methods and the LC-MS method by the Center of Disease Control and Prevention (CDC), together with low variability among the blind duplicates, suggests that both methods reported here would be suitable for the analysis of urinary AM as a biomarker for human exposure of atrazine.