Nonsteroidal anti-inflammatory drugs (NSAIDs) are effective and necessary for the relief of pain and inflammation in patients with arthritis. NSAIDs are however also associated with an increased risk for ulceration in the stomach and in the duodenum, and many NSAID users experience bothersome dyspeptic symptoms during continued NSAID therapy. PPIs like omeprazole, have been shown to heal and to prevent ulcers and dyspeptic symptoms during continued NSAID therapy, and during continued NSAID therapy the prostaglandin analogue, misoprostol, has been shown to reduce the risk for ulcer complications. The COX-2 selective NSAID, rofecoxib, is in comparison with naproxen, a non-selective NSAID, associated with fewer clinically important upper gastrointestinal events. The incidence of myocardial infarctions seems, however, to be lower with naproxen than with rofecoxib, and this is expected to lead to low-dose aspirin use in rofecoxib users at risk for cardiovascular events. Co-administration of the COX-2 selective NSAID, celecoxib, and low-dose aspirin, is associated with the same risk for upper gastrointestinal ulcer complications alone and combined with symptomatic ulcers, as the non-selective NSAIDs, ibuprofen and diclofenac. A proton pump inhibitor (PPI) should be used for healing of NSAID-associated ulcers, and a PPI or misoprostol should be considered for prevention of ulceration in non-selective NSAID users at risk for ulceration. The experience with COX-2 selective NSAIDs is still limited, and it remains to be studied whether subpopulations of COX-2 selective NSAID users will benefit from gastro-duodenal protection. (C) 2001 Lippincott Williams & Wilkins.