The effects of 14 non-steroidal anti-inflammatory drugs (NSAIDs)--naproxen, ibuprofen, mefenamic acid, ketoprofen, indomethacin, fenoprofen, diclofenac sodium, aspirin, salicylic acid, piroxicam, sulindac, fenbufen, flurbiprofen and benzydamine, on the plasma protein binding of thiopental and the clinical consequences of such interactions were studied. Four of them, naproxen, ibuprofen, salicylic acid and aspirin, very significantly decreased the protein binding of thiopental in vitro in human plasma (P < 0.005). Structurally, they were salicylates and propionic acid derivatives among the six classes of NSAIDs studied. The aspirin study demonstrated that the protein-displacing phenomenon was temperature-dependent, and concentration-dependent. Clinically, aspirin administered intravenously resulted in a significant increase in the percentage of plasma free thiopental from 16.01 +/- 3.59% to 22.27 +/- 3.96% (P < 0.001, n = 10) in patients undergoing surgery, and resulted in three of seven patients sleeping again during recovery from thiopental-induced anesthesia. Although the effect of chronic use of NSAIDs before anesthesia is uncertain, studies should be carried out to find out if naproxen, ibuprofen, and aspirin influence the depth of anesthesia, time of recovery and duration of action of thiopental.