Objective: To study the effects of COX-2 on colonic surgical wound healing.
Background: Cyclooxygenase-2 (COX-2) is a key enzyme in gastrointestinal homeostasis. COX-2 inhibitors have been associated with colonic anastomotic leakage.
Methods: Wildtype, COX-2 knockout and COX-2 heterozygous mice were subjected to a model of colonic anastomotic leakage, and were treated with vehicle, diclofenac, or prostaglandin E2 (PGE2), the most important COX-2 product in the intestine. We assessed anastomotic leakage, mortality, angiogenesis, and inflammation. Furthermore, we investigated the association between anastomotic leakage and a human polymorphism of the COX-2 gene resulting in low COX-2 levels.
Results: Diclofenac, a nonsteroidal anti-inflammatory drug inhibiting COX-2, increased anastomotic leakage compared to vehicle-treated mice (100% vs 25%, respectively). Similarly, 92% of COX-2-deficient mice developed anastomotic leakage (P = 0.003) compared to WT. PGE2 partly rescued this severe phenotype because only 46% of PGE2-administered COX-2 knockout mice developed anastomotic leakage (P = 0.02). This may be related to decreased neovascularization, because decreased CD31 staining, indicating less blood vessels, was observed in COX-2 mice (2 vessels/mm vs 6 vessels/mm in controls (P = 0.03)). This effect could partly be reversed by administration of PGE2 to COX-2 mice. No significant differences in inflammation were found. PTGS2-765G>C polymorphism in humans, associated with reduced COX-2 expression, was associated with higher anastomotic leakage rates.
Conclusions: COX-2-induced PGE2 production is essential for intestinal wound healing after colonic surgery, possibly via its effects on angiogenesis. These data emphasize that COX-2 inhibitors should be avoided after colonic surgery, and administration of PGE2 might be favorable for a selection of patients.