1. Paraoxon concentration was estimated by means of inhibition kinetics observed with electric eel acetylcholinesterase (AChE) which was determined by a modified Ellman procedure. In human plasma, paraoxon was stabilized by inactivation of paraoxonase with EDTA and aluminon and by inhibition of butyrylcholinesterase with ethopropazine. Paraoxon (1-50 ng) was recovered at 86+/-1.7% (mean+/-s.e.m.) in ether extracts from 0.5 ml samples of spiked stabilized plasma. It could be stored without loss at - 20 degrees C for at least 1 month. 2. The enzyme-based assay was applied to follow the paraoxon plasma concentrations in three suicidal patients with severe parathion poisoning. In poisoning with excessive doses and initial paraoxon concentrations above 500 nM, therapeutic obidoxime concentrations of approximately 10 microM failed to essentially reactivate erythrocyte AChE in vivo, while reactivatability ex vivo was nearly complete. With the plasma concentrations of paraoxon dropping below 100 nM, however, reactivation by obidoxime became significant. Unexpectedly, paraoxon levels occasionally reincreased during treatment and resulted in re-inhibition of AChE, bearing some resemblance to the Intermediate Syndrome. 3. The paraoxon concentrations measured fitted satisfactorily the values calculated from the kinetic constants previously obtained for AChE inhibition and obidoxime-induced reactivation in vitro. This indicates that diethylphosphoryloxime formation during obidoxime-induced reactivation does not markedly contribute to the re-inhibition of AChE as observed in vitro.