We have developed a unique rat AGML model produced by ischemia/reperfusion plus 0.2% ammonia (I/R.NH3), either treatment which would not induce mucosal injury when used alone. The effects of troxipide and other gastric mucosal defensive drugs were investigated with this I/R.NH3-induced AGML model and other AGML models in rats. The following results were obtained: 1) Like allopurinol, troxipide at 50-200 mg/kg, p.o. dose-dependently prevented I/R.NH3-induced development of AGML and also the ischemia/reperfusion-induced increase of gastric mucosal thiobarbituric acid (TBA)-reactive substances; 2) Troxipide at 10(-6)-10(-4) M, like allopurinol, inhibited concentration-dependently in vitro xanthine oxidase activity in gastric mucosal homogenates; 3) Troxipide at 50-200 mg/kg, p.o. inhibited AGMLs induced by bleeding plus 0.2% ammonia and by 1.0% ammonia alone; and 4) Troxipide and sofalcone were similar in preventing all AGMLs tested and also the increase of mucosal TBA-reactive substances, but somewhat differed from teprenone, cetraxate hydrochloride, azulene plus L-glutamine and sucralfate. These findings suggest that troxipide may inhibit I/R.NH3-induced AGML development by preventing generation of oxygen free radicals and by protecting against mucosal fragility due to reduced energy metabolism from poor blood flow and also against ammonia-induced disruption of the gastric mucosal barrier. Therefore, troxipide may be highly effective for various AGMLs with multifactor involvement.