The elderly rheumatic patients and 7 healthy young persons received naproxen (Naprosyn, Syntex) 500 mg orally twice a day for 4 weeks. The serum concentrations were determined using mass fragmentography. After an initial 1,000-mg dose, no significant differences were found between the two groups in peak serum concentration, time to peak serum concentration, area below the serum concentration-time curve, volume of distribution, elimination half-life, or total body clearance of naproxen. At steady state, the median total through naproxen concentration was 50.5 mg/l in the elderly and 62.7 in the young (p = 0.08); the unbound concentration was 58 micrograms/l and 44 micrograms/l, respectively (p = 0.06). There was a significant inverse correlation between serum albumin and the free fraction of naproxen (R = -0.58, p = 0.01). The hepatic extraction ratio of naproxen is relatively low and it is suggested that the reduced protein binding in the elderly may conceal the age-related reduction in cellular activity. An estimated value of intrinsic clearance was reduced by 37% in the elderly patients. It is suggested to start naproxen at the lower end of its dose range in the elderly.