It appears that selective Cox-2 inhibitors do not affect the gastro duodenal mucosa whilst having antiinflammatory and analgesic efficacy similar to non-selective NSAIDs. Two broad categories of drugs are Cox-2 selective: coxibs and a number of pre-existing NSAIDs retrospectively found to have selectivity. Cox-2 inhibitors cause less dyspepsia than NSAIDs. They spare gastrointestinal mucosal generation of prostaglandins (PGs) and PGdependant bicarbonate secretion. Coxibs cause no acute mucosal injury in endoscopic studies and serendipitous Cox-2 inhibitors generally cause less acute injury than non -selective NSAIDs or placebo. Both celecoxib and rofecoxib have been associated with a substantial reduction in endoscopic ulcers compared to NSAID comparators. In the VIGOR study all upper GI events were reduced from 4.5 per 100 patient years to 2.1 per 100 patient years with supra-therapeutic doses of rofecoxib compared with naproxen. In the CLASS study, over a period of 3 days to 6 months, incidence of ulcer complications was 0.76% with celecoxib and 1.45% for ibuprofen or diclofenac. The less substantial reduction in events in the CLASS study compared with the VIGOR may be due, at least in part, to the fact that 21% of the patients were also on low dose aspirin. However it is premature to say that the benefit of Cox-2 inhibitors is lost in patients taking aspirin. There is continuing debate on the role of Cox-2 inhibitors in patients who have other risk factors for complicated ulcer disease e.g. patients who are elderly, on aspirin or corticosteroids, have a previous ulcer or have H. pylori infection.