Where is the evidence that cyclooxygenase inhibition is the primary cause of nonsteroidal anti-inflammatory drug (NSAID)-induced gastrointestinal injury? Topical injury revisited.

In this commentary, we take a critical look at the concept that the gastrointestinal (GI) side-effects of nonsteroidal anti-inflammatory drugs (NSAIDs) are due to the ability of these drugs to inhibit cyclooxygenase-1 (COX-1) that is constitutively expressed in the GI mucosa. Indeed, development of the new "super aspirins," such as Celebrex and Vioxx, that selectively inhibit the inducible COX-2, expressed in areas of inflammation, is a direct outgrowth of this concept. We discuss evidence from both the laboratory and the clinic that appears to be inconsistent with the above concept, and cite a number of examples where the depletion of mucosal prostaglandin levels and the development of GI injury can be dissociated. Instead, we revisit the possibility that NSAID-induced GI side-effects are mostly due to the ability of these drugs to topically injure the GI mucosa. We devote the remainder of the commentary to presenting evidence from our and other laboratories that NSAIDs can directly attenuate the surface hydrophobic barrier of the GI mucosa due to their ability to bind to zwitterionic phospholipids, and that even systemically administered NSAIDs that are secreted into the bile may induce GI ulceration and/or bleeding due to phospholipid interactions and the development of topical mucosal injury.