OBJECTIVE To test the hypothesis that unbound concentrations of naproxen in synovial fluid (SF) and plasma (P) are equal over a drug dosage interval at steady state or after a single dose of drug. The relationship between plasma and SF concentrations of naproxen, respectively, and prostaglandin concentrations were also examined. METHODS Paired, sequential, total, and unbound naproxen concentrations were determined in plasma and SF in 2 groups of 6 patients. A single dose group was given naproxen 500 mg. The chronic dose group was given 500 mg bd for 7 days before collection of blood and SF samples. The effect of naproxen on prostanoid production by clotting whole blood (thromboxane B2, TXB2) and in SF (PGE2, 6-keto-PGF1 alpha) was determined by radioimmunoassay. RESULTS Average area under the curve (AUC) of unbound (U) naproxen concentrations against time in plasma and SF were the same over a dosage interval at steady state (ratio AUCU,SF/AUCU,P, 1.12 +/- 0.18; p = 0.108), but not after a single acute dose (AUCU,SF/AUCU,P, 1.34 +/- 0.32; p = 0.044). Data from the single dose study revealed that the mean (+/- SD) of the concentrations required for 50% inhibition (EC50) of platelet derived TXB2 by total naproxen was 7.7 +/- 4.4 micrograms/ml (n = 5) and for unbound drug 25.4 +/- 22.0 ng/ml (n = 5). SF prostanoid concentrations after both acute and chronic dosing were low, as expected, but temporal and dose relationships of prostanoid concentrations with SF naproxen could not be discerned. However, this may reflect study design. CONCLUSIONS The AUC of unbound naproxen in SF and plasma were similar at steady state. Plasma concentrations correlated with inhibition of TXB2 generation by platelets. There was sustained depression of PG concentrations in SF beyond the time suggested by plasma drug concentrations.