To investigate the hypothesis that reactive oxygen metabolites are important in the pathophysiology of duodenal ulcer disease, their production by duodenal mucosal biopsy specimens was measured using luminol and lucigenin amplified chemiluminescence. Luminol chemiluminescence, expressed as background corrected median photon emission/mg/min x 10(3) (95% confidence intervals), was increased in duodenal inflammation as assessed macroscopically: ulcers 20.3 (4.8 to 51.3), n = 29; severe duodenitis 13.9 (6.6 to 75.3), n = 16; mild duodenitis 0.0 (-0.5 to 0.8), n = 56; controls -0.8 (-1.3 to -0.1), n = 41; p = 0.0001, Kruskal-Wallis) and microscopically: severe 17.0 (9.3 to 51.3), n = 12; moderate 0.3 (-2.8 to 5.8), n = 17; mild -0.1 (-1.8 to 1.0), n = 17; controls -0.8 (-1.6 to 0.0), n = 15; (p = 0.0001). Luminol chemiluminescence was directly related to both the macroscopic and microscopic severity of duodenal damage (Spearman's R = + 0.53, + 0.55 respectively, both p = 0.0001), to histochemical assessment (myeloperoxidase activity) of neutrophil infiltration (R = + 0.63; p = 0.04), and to lucigenin chemiluminescence (R = + 0.56, p = 0.0002). Luminol chemiluminescence was inhibited by sodium azide (-80%), catalase (-73%), and dimethyl sulphoxide (-24%). Superoxide dismutase inhibited lucigenin more than luminol dependent chemiluminescence (-61% and -7% respectively, p < 0.05). Within disease groups, Helicobacter pylori antral infection was associated with increased duodenal chemiluminescence, whereas smoking, alcohol, and use of NSAIDs or H2 blockers had no influence. Their disease related generation in duodenal mucosa supports a role for reactive oxygen metabolites in the pathogenesis of duodenitis and duodenal ulcer. These metabolites might include superoxide, hydrogen peroxide, hydroxyl, and products of myeloperoxidase activity.