Selective inhibitors of the "inducible" isoform of cyclooxygenase (COX-2) have been suggested to be effective analgesic and anti-inflammatory drugs while sparing the gastrointestinal (GI) tract of injury. There is some experimental and early clinical evidence to support this hypothesis. However, some important questions remain regarding the utility of selective COX-2 inhibitors. For example, estimates of the selectivity of COX-2 inhibitors based on in vitro studies are likely to be poor predictors of selectivity in vivo. Efficacy with selective blockade of COX-2 may be inferior to that achieved with combined inhibition of COX-1 and COX-2. Furthermore, in situations in which there is inflammation or ulceration in the GI tract, COX-2 produces prostaglandins that are essential for repair. In these circumstances, inhibition of COX-2 leads to delay of ulcer healing and exacerbation of inflammation. Some caution should therefore be exercised before the theory is fully accepted that selective COX-2 inhibitors are effective anti-inflammatory drugs that spare the GI tract of injury.