The pathogenesis of nonsteroidal anti-inflammatory drug (NSAID) enteropathy is a complex process involving the uncoupling of mitochondrial oxidative phosphorylation and inhibition of cyclooxygenase (COX). Rofecoxib, a selective inhibitor of COX-2, has shown less gastric damage, but the same beneficial effect is not clear in the case of the small bowel. Fifty-seven male Wistar rats (250-350 g) were divided into three groups (N=19 each) to evaluate the effect of this NSAID on the rat intestine. The groups received 2.5 mg/kg rofecoxib, 7.5 mg/kg indomethacin or water with 5% DMSO (control) given as a single dose by gavage 24 h before the beginning of the experiment. A macroscopic score was used to quantify intestinal lesions and intestinal permeability was measured using [51Cr]-ethylenediaminetetraacetic acid ([51Cr]-EDTA). The extent of intestinal lesion, indicated by a macroscopic score, was significantly lower when rofecoxib was administered compared to indomethacin (rofecoxib=0.0 vs indomethacin=63.6 +/- 25.9; P<0.05) and did not differ from control. The intestinal permeability to [51Cr]-EDTA was significantly increased after indomethacin (control=1.82 +/- 0.4 vs indomethacin=9.12 +/- 0.8%; P<0.0001), but not after rofecoxib, whose effect did not differ significantly from control (control=1.82 +/- 0.4 vs rofecoxib=2.17 +/- 0.4%; ns), but was significantly different from indomethacin (indomethacin=9.12 +/- 0.8 vs rofecoxib=2.17 +/- 0.4%; P<0.001). In conclusion, the present data show that rofecoxib is safer than indomethacin in rats because it does not induce macroscopic intestinal damage or increased intestinal permeability.