Abstract-: American soldiers involved in friendly fire accidents during the 1991 Gulf War were injured with depleted-uranium-containing fragments or possibly exposed to depleted uranium via other routes such as inhalation, ingestion, and/or wound contamination. To evaluate the presence of depleted uranium in these soldiers eight years later, the uranium concentration and depleted uranium content of urine samples were determined by inductively coupled plasma mass spectrometry in (a) depleted uranium exposed soldiers with embedded shrapnel, (b) depleted uranium exposed soldiers with no shrapnel, and (c) a reference group of deployed soldiers not involved in the friendly fire incidents. Uranium isotopic ratios measured in many urine samples injected directly into the inductively coupled plasma mass spectrometer and analyzed at a mass resolution m/Δm of 300 appeared enriched in 235U with respect to natural abundance (0.72%) due to the presence of an interference of a polyatomic molecule of mass 234.81 amu that was resolved at a mass resolution m/Δm of 4,000. The 235U abundance measured on uranium separated from these urines by anion exchange chromatography was clearly natural or depleted. Urine uranium concentrations of soldiers with shrapnel were higher than those of the two other groups, and 16 out of 17 soldiers with shrapnel had detectable depleted uranium in their urine. In depleted uranium exposed soldiers with no shrapnel, depleted uranium was detected in urine samples of 10 out of 28 soldiers. The median uranium concentration of urines with depleted uranium from soldiers without shrapnel was significantly higher than in urines with no depleted uranium, though substantial overlap in urine uranium concentrations existed between the two groups. Accordingly, assessment of depleted uranium exposure using urine must rely on uranium isotopic analyses, since urine uranium concentration is not an unequivocal indicator of depleted uranium presence in soldiers with no embedded shrapnel.