Prevention of Intra-abdominal Adhesions Using the Antiangiogenic COX-2 Inhibitor Celecoxib

To the Editor:

We read, with avid attention and close interest, the carefully performed study of Greene et al,¹ demonstrating that perioperative COX-2 inhibitors may attenuate adhesion formation in an experimental murine model. This seems entirely in keeping with the known antiangiogenic effects of these drugs and the up-regulation of COX-2 that is known to occur in adhesion-associated tissues and, especially, fibroblasts. However, we are somewhat concerned that the authors did not extrapolate further from their realizations that adhesion formation is “like all types of wound healing” and that “it may aid in healing” and include formal assessment of surgical wound healing in their protocol. This omission is particularly apparent given that the authors admit awareness of the fact that other antiangiogenic agents have been previously precluded from further investigations as anti-adhesion agents because oftheir effects on wound healing among other problems. Additionally, there exists a considerable body of literature that suggests that COX-2 may function more in a pro-restorative role after tissue injury than in a purely pro-inflammatory capacity. In particular, these anti-inflammatory agents in general (and COX-2 inhibitors in particular) are known to delay the healing of gastric ulcers² as well as intestinal mucosal injuries³ and bone fractures⁴ and have also been shown (by our own group⁵ and others⁶) to markedly affect the integrity of colonic anastomoses in experimental settings similar to that used in this current study. Furthermore, this specific area of concern regarding their safety can be expanded to include the authors other contentions that “celecoxib and rofecoxib are known to be safe even when taken in high doses chronically” and that “clinical trials (now) need to be performed to confirm the (their) beneficial properties in reducing intra-abdominal adhesions.” Unfortunately, it has been recently realized that exactly the opposite is true, and the considerable cardiovascular toxicities induced by these agents have led to the withdrawal of rofecoxib and labeling precautions being ascribed to other related compounds.^7,8

In defense of the authors, however, it may well be that (rather than selective referencing or incomplete background research) their study has been superseded by clinical and experimental realizations that have occurred sometime between its initial submission and entry in to the public domain (a consideration supported by their most recent reference listed being from 2003). Acknowledgment of this lag-time in the section devoted currently to corresponding author details may go some way toward facilitating a reader's appreciation of the current relevance of experimental findings. However, in cases such as this, when manuscripts are, perhaps, more significantly undermined but by their lack of an up-to-date discussion of the clinical relevance of their hypothesis than by their study design or analytical performance, it would seem prudent allow the authors a late right of review of their manuscript or, at least, a prefacing statement to avoid misleading conclusions to be transmitted to the casual reader.

Ronan A. Cahill, AFRCSI
Department of Surgery
Cork University Hospital
Cork, Ireland
rcahill@rcsi.ie