We evaluated the effect of in vivo and in vitro administration of nitro-containing and nitro-deficient L-arginine-derived nitric oxide (NO) synthase inhibitors on the measurement of NO in plasma, urine and HEPES buffered physiologic salt solution (PSS) by ozone chemiluminescence and by the modified Griess reaction. In vivo administration of 1, 5, 25, 40 or 50 mg/kg of NG-nitro-L-or D-arginine methyl ester (LNAME, DNAME), NG-nitro-L-arginine (LNA) or aminoguanidine (AG) to rats and mice increased NO in urine and plasma as determined by chemiluminescence using 2.3% vanadium chloride in 2N HCI at 100 degrees C as the redox reagent. In vivo administration of 1 and 10 mg/kg/day of NG-imino-ethyl-L-ornithine (LNIO) or 3 amino-1,2,4 triazine (AT) reduced plasma and urine NO. Addition of LNAME, DNAME and LNA (100 nM to 1 mM) to the redox solution produced a concentration response curve for NO in the chemiluminescence assay similar to that produced by standard solutions of sodium nitrite and nitrate. LNMMA produced a small NO signal but only at concentrations equal to or exceeding 0.1 mM. LNIO, AT and AG did not give any NO signal even at concentrations exceeding 1 mM. Conversion of plasma or urine nitrate to nitrite with cadmium gave elevated values of plasma nitrite by the Greiss assay when LNAME or LNA was the NO synthase inhibitor. We conclude that in vivo and in vitro use of LNAME and LNA and in vivo use of high doses of aminoguanidine interfere with the assay of NO2- and NO3- with the modified Griess reaction and with chemiluminescence. We suggest that LNAME and LNA not be used in vivo or in vitro when total RNI is measured with these assays.