Several authors have described the occurrence of N-nitrosodimethylamine (NDMA) in body fluids (e.g., blood and urine) and have interpreted this finding as an indication of endogenous formation of NDMA. Controlled excretion studies as well as careful control of artefacts showed, however, that, under normal conditions, NDMA formation in vivo cannot be monitored directly in urine due to a high metabolic conversion rate (more than 99.9%). Our own experiments showed an increased excretion rate (up to 2.4%) when ethanol was administered simultaneously. This model was used in experiments to monitor in-vivo formation of NDMA. Amidopyrine, a compound that is easily nitrosated, was administered as a single oral dose of 500 mg to volunteers. With ingestion of 20-30 g ethanol NDMA could be detected in urine. Negative control experiments indicate that the appearance of NDMA in urine derives from in-vivo nitrosation of the drug. Between 0.5 and 10 micrograms NDMA were excreted within 8 h, and excretion was influenced by salivary nitrite concentrations, which ranged from 5-220 mg/L. By comparison with our earlier excretion studies in humans, it can be assumed that only 1-2% of endogenously formed N-nitrosamine was found in urine. To our knowledge, this is the first time that in-vivo formation of NDMA has been shown directly to occur in humans.