Forty-five healthy men aged 21-34 years took part in a double-blind, parallel-group, placebo-controlled study of the effects of 28 days' treatment with lornoxicam 4 mg twice daily or indomethacin 50 mg twice daily on faecal blood loss and the endoscopic appearances of gastric and duodenal mucosa. After an initial endoscopic examination, subjects received, intravenously, on day 0, autologous erythrocytes labelled with 51Cr. Complete daily faecal collections were then made from days 6-12, 20-26 and 34-40. The drug treatments or placebo were given from days 13-41. Faecal blood loss was calculated from 51Cr-specific activity of blood and faeces. Endoscopy was repeated 4-8 hours after the last dose of medication; mucosal appearance was graded on a 5-point scale. Lornoxicam caused no more adverse events than placebo; indomethacin caused more indigestion and central nervous system effects, and one subject in this group was withdrawn from the study. Median total blood losses during the pre-treatment and the second and fourth weeks of treatment were respectively 3.33, 3.95 and 5.71 ml for lornoxicam, 2.87, 7.04 and 7.75 ml for indomethacin, and 4.55, 3.64 and 4.13 ml for placebo. Differences between treatments were not statistically significant (P = 0.081 for second week of treatment, P = 0.383 for fourth week of treatment; Kruskal-Wallis test). The effect of chlortenoxicam on faecal blood loss in this study was thus intermediate between placebo and indomethacin, but within- and between-subject variability was such that the differences were not statistically significant. Endoscopic findings were normal in most subjects before and after all treatments, but indomethacin was associated with a slightly greater deterioration in endoscopic score and was the only treatment associated with Grade 3 appearance (in a single patient) in post-treatment endoscopy.