Five different preparations of diclofenac-suppositories (A, B, C, D, E) are investigated for in vitro liberation (modified paddle method) and for in vivo bioavailability in man (8-12 subjects) in a crossover design. In vitro, the time at which 63.2% of the drug are liberated (tL) is 14, 20, 25, 14 and 3.5 min for the preparations A, B, C, D and E, respectively. The steady state concentrations from the preparations B and C are 51 and 11%, respectively, and lower than for the others. In vivo, the time of the concentration maximum (tmax) in min is (mean +/- S mean): A = 68 +/- 18, B = 72 +/- 9, C = 120 +/- 22, D = 42 +/- 3, E = 24 +/- 0.5. The concentration maximum (cmax) in mumol . l-1 at tmax is (mean +/- S mean): A = 5.9 +/- 0.8, B = 3.9 +/- 0.3, C = 3.1 +/- 0.4, D = 5.7 +/- 0.6, and E = 5.5 +/- 0.8. The area under the curve (AUC) for all the preparations has found to be between 8.7 and 10,6 mumol . h . l-1. There is a significant correlation between the in vitro parameter tL and the in vivo data of tmax and cmax, respectively. In consequence, the invasion behaviour in vivo can be derived from in vitro data for drugs with similar good physicochemical properties as diclofenac-Na.