BACKGROUND: Nitric oxide (NO) produced in excess by the inflamed human colon is generally considered a pathway of mucosal damage. In an attempt to quantify colonic mucosal production of NO in various forms of colitis we performed 'steady-state' gas perfusion of whole colon in 11 patients with ulcerative colitis, 10 patients with collagenous colitis and 20 controls with uninflamed mucosa. METHODS: The tip of a Teflon tube was placed in the caecum during colonoscopy. Subsequently, argon was infused at a constant rate for 70-180 min. Argon and NO in gas sampled from the rectum were measured by neutron activation analysis and the chemiluminescence technique, respectively. RESULTS: The use of argon as a marker of colonic NO output was justified by complete recovery (96%+/-2; mean +/- s(x); n = 5) of argon in gas collected from the rectum and a constant output of NO at varying perfusion rates (25, 50 and 75 ml/min coefficient of variation 21%; n = 6). In patients with ulcerative colitis, colonic output of NO was 10-fold higher (P < 0.001) than in controls and positively correlated (P < 0.01) to indices of disease activity. In patients with collagenous colitis, colonic output of NO was 50-fold higher (P < 0.01) than in controls during periods with watery diarrhoea (n = 6), but within the range observed in ulcerative colitis in the absence of diarrhoea (n = 4). CONCLUSIONS: Argon gas perfusion of whole colon using chemiluminescence technique for measurement of NO is a reliable method for quantification of colonic mucosal NO production. Increased colonic production of NO in collagenous colitis, which exceeds the output observed even in extensive ulcerative colitis, militates against the theory that NO per se is a cause of mucosal injury.