This double-blind study assessed the acute development of NSAID-associated gastroduodenal (GD) damage and its prevention by misoprostol. Patients requiring chronic NSAID therapy were stratified into two groups depending on initial endoscopic appearance, Group I: normal (n = 223); Group II: non-ulcer lesions (n = 78). After 2 weeks of therapy with NSAID and either misoprostol 400-800 micrograms daily or placebo the incidence of severe mucosal damage (including ulcers) was significantly reduced by misoprostol (odds ratio; 95% CI). Group I: 4.52; 1.94, 10.51 (P = 0.018); Group II: 10.93; 1.09, 109.60 (P = 0.014); Groups I and II combined: 5.95; 3.23, 10.94 (P = 0.0003). Misoprostol exerted a significant protective effect against progression of minor to severe damage in Group II (P < 0.001). Endoscopic findings did not correlate significantly with gastrointestinal symptoms and misoprostol did not interfere with the NSAID efficacy. Significant GD damage occurs early in the course of NSAID treatment and misoprostol significantly reduces the incidence of such damage.