OBJECTIVES--To analyse whether the intraindividual variability of the bioavailability of oral methotrexate in rheumatoid arthritis (RA) is as high as the interindividual variability and whether the bioavailability testing can be simplified. METHODS--Fifteen mg methotrexate was given orally after an overnight fast to 10 patients with RA on two occasions, one week (n = 4) or two years (n = 6) apart, respectively. Plasma samples were taken at specific time intervals and analysed by fluorescence polarisation immunoassay. The areas under the plasma concentration versus time curves from 0 to infinity (AUC) were calculated after fitting to a two-compartment extravascular model with lag time. RESULTS--The interindividual variability of the AUCs showed a more than five fold range from 77 to 471 min x mumol/l. In contrast, the intraindividual differences of AUCs between the two visits ranged from 3 to 100 min x mumol/l, reflecting a 3-30% change in nine of 10 patients. Close correlations were found between the AUCs and the plasma concentrations with a most pronounced correlation eight hours after methotrexate intake (r = 0.975; p < 0.001). CONCLUSIONS--The high interindividual variability of the AUCs was confirmed. Conversely, only a modest intraindividual variability was disclosed. Plasma sample analysis at a single time point four to eight hours after methotrexate application could simplify estimation of the bioavailability.