Most animal models of nonsteroidal anti-inflammatory drug (NSAID) induced gastric injury are characterized by acute, superficial erosions in the corpus region, whereas the clinically significant injury in man is the deep, antral ulcer. The purpose of this study was to characterize a model of NSAID-induced antral ulceration that more closely resembles the type of damage seen in man. Rabbits received indomethacin subcutaneously every 12 h. The progression of ulcer formation was followed by killing groups of animals after one to seven doses of indomethacin. The dose dependency of ulcer formation was assessed by giving indomethacin at doses of 1 to 20 mg/kg. Healing of antral ulcers was determined by examining the stomach at various times after administering the seventh dose of indomethacin (20 mg/kg). The effects of prophylactic treatment with misoprostol or ranitidine on ulcer formation were assessed. Indomethacin administration initially produced superficial erosions in the corpus and antrum, but with time, ulcers became apparent in the antrum. The formation of these ulcers was dependent upon the number of times indomethacin was administered and the dose. Similar ulcers could be induced with a second NSAID, diclofenac. Misoprostol treatment resulted in a significant reduction in the extent of indomethacin-induced antral ulceration, but ranitidine had no effect. Antral ulcers healed progressively following cessation of indomethacin administration and were almost completely resolved by 108 h after the final dose of indomethacin. These results demonstrate that subcutaneous NSAID administration to rabbits is a simple and reproducible method for producing ulcers that bear striking macroscopic resemblance to NSAID-induced antral ulcers in man.(ABSTRACT TRUNCATED AT 250 WORDS)