The purpose of this study was to develop an acute animal model of reflux esophagitis, which would be suitable to induce the esophageal damage caused by gastric acid reflux, thus mimicking the esophageal injury of human gastroesophageal reflux disease (GERD). Global research indicates that GERD is rapidly increasing among the world's population. NSAIDs are known to induce gastrointestinal damage and low doses of aspirin (ASA) have been shown to increase the incidences of GERD in humans. Gastric acid and pepsin secretion and enhanced COX-2 expression were implicated in the pathogenesis of reflux esophagitis, but the effect of selective COX-2 inhibitors against lesions induced by the reflux of gastric acid content into esophagus has not been thoroughly studied. Here, we compared the effect of aspirin (ASA) and so called "safe" nitric oxide (NO) derivative of ASA with those of non-selective and selective cyclooxygenase (COX)-1 and COX-2 in rat model of reflux esophagitis. Reflux esophagitis was induced in anesthetized rats by ligating the pylorus and limiting ridge transitional region between the forestomach and the corpus of stomach. Subsequently, the total gastric reservoir to store gastric juice was greatly diminished, resulting in the reflux of this juice into the esophagus. Rats with esophagitis received intragastric (i.g.) pretreatment either with: 1) vehicle (saline), 2) ASA or NO-ASA (100 mg/kg); 3) the non-selective COX inhibitor, indomethacin (5 mg/kg); 4) the selective COX-1 inhibitor, SC-560 (10 mg/kg), and 5) the selective COX-2 inhibitor, celecoxib (5 mg/kg). In a separate series of rats with reflux oesophagitis, the efficacy of ASA combined with a donor of NO, glyceryl trinitrate (GTN; 10 mg/kg i.g.) to prevent esophageal mucosal injury was investigated. Four hours after induction of esophagitis the gross mucosal damage was graded with a macroscopic lesion index (LI) from 0-6. The esophageal blood flow (EBF) was determined by H₂-gas clearance technique, the oesophageal mucosal and blood samples were collected for histology and analysis of the RT-PCR expression and release of proinflammatory cytokines IL-1β, TNF-α and IL-6 using specific ELISA. The exposure of the esophagus to reflux of gastric acid time-dependently increased the esophageal LI and morphologic damage, and decreased EBF with the most significant changes observed at 4 hrs after the ligation procedure. The pretreatment with native ASA in the dose that suppressed the generation of mucosal PGE₂, enhanced gross and histologic esophageal damage and produced a significant fall in EBF. NO-ASA or ASA coupled with GTN counteracted the aggravation of the damage and accompanying fall in EBF when compared with native ASA applied alone to rats with esophagitis. The proinflammatory cytokines IL-1β and TNF-α were overexpressed in rats with esophagitis and those pretreated with ASA but this effect was significantly attenuated by NO-ASA. Plasma IL-1β, TNF-α and IL-6 were negligible in the intact rats but significantly increased in those with esophagitis, with this effect being further enhanced by non-selective (indomethacin) and selective (SC-560, celecoxib) COX-1 and COX-2 inhibitors. We conclude that conventional NSAID such as aspirin augments esophagitis, while NO-ASA exerts the beneficial protective effect against reflux esophagitis via the enhancement of esophageal microcirculation due to NO release and an inhibitory effect on expression and release of pro-inflammatory cytokines.