Identified studies suggest that cox-II inhibitors are associated with a slightly reduced risk of gastrointestinal adverse events compared with NSAIDs. However, the reliability of the conclusion is limited by low power for detecting adverse effects. In many cases, adequate power was achieved by pooling results from different exposure groups (receiving different NSAIDs or different doses). This prevents direct comparison of adverse effects of particular cox-II inhibitors versus particular NSAIDs. Results may not be applicable to people at high risk of gastrointestinal adverse effects. Most studies excluded people with a recent history of serious gastrointestinal disease. Similarly, we found insufficient evidence to compare gastrointestinal adverse effects of cox-II inhibitors versus NSAIDs in high-risk patients such as those chronically using corticosteroids or those of older age. Concomitant use of H2-antagonists and/or proton pump inhibitors was not reported (in the NICE review) or not allowed (in the subsequent RCTs). We cannot, therefore, reliably answer the clinically important question of whether cox-II inhibitors reduce gastrointestinal adverse effects compared with combined therapy with NSAIDs plus H2 antagonists or proton pump inhibitors.