The role of arachidonic acid metabolites and the mode of action of 5-aminosalicylic acid, the active moiety of sulphasalazine and disodium axodisalicylate, in ulcerative colitis remain obscure. Therefore, experiments were performed in which the effects of medication on immunoreactive prostaglandin (PG) E2 concentrations in free faecal water were assessed using the equilibrium in vivo dialysis of faeces. Colonic PGE2 concentrations in patients with active ulcerative colitis (n = 11) ranged from 2035-18 000 pg/ml to be compared with a range of 103-188 pg/ml in healthy volunteers (n = 10; p < 0.001). In all healthy volunteers PGE2 concentrations decreased slightly (p < 0.05) after disodium azodisalicylate intake 2 g/day, whereas low dose disodium azodisalicylate (0.25 g/day) caused no change. In patients with ulcerative colitis in complete clinical, sigmoidoscopic, and histologic remission withdrawal of sulphasalazine (2 g/day; n = 6) increased PGE2 concentrations to values above normal levels (p < 0.05) which returned to pretrial values (p < 0.05) on disodium azodisalicylate (2 g/day; n = 7). In conclusion, increased PGE2 in free faecal water indicates an abnormality in the colonic mucosa, even in the absence of conventional signs of inflammation. We could not confirm the hypothesis that sulphasalazine and 5-aminosalicylic acid exert their therapeutic effect through promotion of endogenous cytoprotective prostaglandins. In contrast, the observation that raised PGE2 concentrations were normalised by sodium azodisalicylate in patients with inactive ulcerative colitis suggests that subclinical disease activity was decreased by 5-aminosalicylic acid.