AIM: To determine whether misoprostol promotes the healing of non-steroidal anti-inflammatory drug-induced gastroduodenal lesions in a human experimental model. METHODS: Mucosal damage and healing of mucosal biopsy sites were assessed endoscopically in 10 healthy, Helicobacter pylori-negative volunteers with a normal initial endoscopy: they were enrolled in a double-blind, double-dummy, placebo-controlled cross-over study. They received 2-week courses of misoprostol (200 micrograms b.d.) or placebo; a water-soluble non-steroidal antiinflammatory drug diclofenac 50 mg t.d.s., was given during the second week of each dosage regimen after three endoscopic biopsies had been taken from each of the duodenum, antrum and corpus. RESULTS: The number of unhealed biopsy sites was not different after misoprostol or placebo, although the number of healed biopsy sites was greater in the corpus and duodenum than in the antrum. Misoprostol did not prevent the appearance of diclofenac-induced erosions and petechiae. Epigastric discomfort was related to the intake of diclofenac and was reduced by misoprostol. Bloating and flatulence occurred more frequently with misoprostol alone and with misoprostol plus diclofenac, than with placebo alone or placebo plus diclofenac. CONCLUSION: Misoprostol does not prevent new mucosal lesions induced by diclofenac in healthy volunteers and it does not accelerate the healing of the biopsy sites. Misoprostol decreases the frequency of diclofenac-induced epigastric discomfort, but it increases gas bloating and flatulence.