OBJECTIVE To analyse the influence of age and cytochrome P450 (CYP) 2C9 genotype on the steady-state disposition of the standard NSAID diclofenac and the new COX-2 selective inhibitor celecoxib, both of which are metabolised by the polymorphically expressed CYP2C9. DESIGN Double-blind randomised crossover study under steady-state conditions. SUBJECTS 12 young (age 32 +/- 5 years, bodyweight 71 +/- 12kg; mean +/- SD) and 12 elderly (68 +/- 2 years, 82 +/- 15kg) healthy, drug-free, nonsmoking Caucasians of both sexes. METHODS All subjects received oral celecoxib (200mg twice daily) and diclofenac (75mg twice daily) for 15 days separated by a drug-free interval of at least 3 weeks. Following the last morning dose, multiple blood samples were taken for 25 hours. Concentrations of celecoxib and diclofenac were measured by specific and sensitive high performance liquid chromatography. Identification of CYP2C9 genotype was performed by genomic DNA sequencing. Pharmacokinetic parameters for total and unbound drugs were individually analysed by noncompartmental techniques. RESULTS For diclofenac, area under the concentration-time curve over the dosage interval (AUC(tau)) was larger in young subjects (3.2 +/- 1.0 mg * h/L) than in older individuals (2.4 +/- 0.4 mg * h/L; p < 0.05). As the terminal half-life (t((1/2)Z)) was very similar in both groups (3.9 +/- 4.4 vs 3.5 +/- 3.3 hours), either less complete absorption in the elderly or their higher bodyweight could account for the difference. For celecoxib, AUC(tau) (5.8 +/- 1.7 vs 5.6 +/- 2.3 mg * h/L) and t((1/2)z) (11.8 +/- 8.7 vs 11.2 +/- 2.9 hours) were almost identical in young and older subjects. Plasma protein binding of both NSAIDs was unaffected by age, and apparent oral clearances for unbound drugs were not different between the two groups of healthy subjects. When considering the genotype of all individuals (CYP2C9*1/*1, n = 10; CYP2C9*1/*2, n = 6; CYP2C9*2/*2, n = 2; CYP2C9*1/*3, n = 4; CYP2C9*3/*3, n = 1), no association with any pharmacokinetic parameter of either drug was apparent. Moreover, there was no significant correlation between the AUC values of celecoxib and diclofenac. CONCLUSIONS Age and CYP2C9 genotype do not significantly affect the steady-state disposition of celecoxib and diclofenac. This would indicate that both drugs need no dosage reduction in the elderly (at least up to 75 years) and that, besides CYP2C9, additional CYP species contribute to the elimination of both agents.