Double-blind dose-response studies of aspirin, phenylbutazone, indomethacin, ibuprofen, naproxen, piroxicam, diclofenac, diflunisal, and tolfenamic acid were reviewed, using pain and a combined effect measure as outcome variables. The increase in effect obtained with the highest doses of the drugs was small, and as a general rule, NSAIDs should be used at the lower end of the recommended dose ranges.