A comparison has been made between accelerator mass spectrometry (AMS) analysis and liquid scintillation counting (LSC) of plasma, urine and faecal samples containing 14C-labelled drugs. In an in vitro study in which human plasma was spiked (the term spiked is used in Section 2.6) with 14C-Fluconazole (14C-FL) over a concentration range of 0.1-2.5 dpm/ml, a correlation coefficient of 0.999 was determined for AMS analysis versus extrapolated LSC data. No significant day to day (or inter-day)variation was seen (P < 0.05 by ANOVA). Coefficients of variation for these analyses ranged from 2.68 to 6.50%. In vivo studies in which rats were given a high (11.5 microCi/kg) or low (18.1 nCi/kg) radioactive dose (to model an exposure of 0.9 microSievert to man) of 14C-Fluticasone propionate(14C-FP) showed that there was also a good correspondence between AMS and LSC data. A mass balance study in a single the faeces by 96 h; less than 1% of the administered dose was excreted in the urine. The limit of reliable measurement of drug related material, above background concentrations, by AMS analysis in this study was approximately 0.1 dpm/ml for plasma, 0.01 dpm/ml for urine without any sample extraction or concentration and 0.01 dpm/ml for faecal extracts. The data reported here demonstrate that AMS is an ultrasensitive and reliable method for analysing 14C-labelled drugs in human and animal body fluids.