The absorption of flurbiprofen through human buccal mucosa was studied after 25 mL of a 2.5-mg/mL solution (pH 4) of the drug in a cosolvent mixture (ethanol 95%:glycerin:propylene glycol:0.3 M sodium acetate buffer, 10:40:30:20) was held and circulated in the mouth for 5 min in a "buccal absorption test." The results were compared with those obtained after oral administration of 25 mL of a solution (pH 7) of sodium flurbiprofen having the same concentration. Twelve subjects participated in the crossover study. After the buccal treatment, mean Cmax and Tmax values were 0.751 micrograms/mL and 41 min, respectively. Average Cmax and Tmax values after the oral treatment were 10.8 micrograms/mL and 32 min, respectively. Mean dose-corrected AUCs were 0.0854 and 0.811 (micrograms.h/mL)/mg for the buccal and the oral treatments, respectively. The absorption kinetics after the buccal treatment were evaluated using the Exact Loo-Riegelman Method (ELRM). Buccal plasma flurbiprofen concentration-time data for 11 subjects were very well fitted by reconstructed curves using ka and the lag-time obtained from ELRM analysis of the buccal data and the disposition parameters obtained from the oral data. These results strongly support the concept of intrasubject constancy of flurbiprofen disposition parameters. Analysis, by ELRM, of the plasma concentration-time data obtained after the buccal treatment indicated first-order absorption, with a mean ka value of 3.9 +/- 2.2 h-1. This value was significantly different (0.05 greater than p greater than 0.02) from the ka after oral treatment (7.89 +/- 5.2 h-1), obtained from the triexponential fitting of the oral plasma data.(ABSTRACT TRUNCATED AT 250 WORDS)