Inhibition of tumor necrosis fact (TNFalpha) is of potential benefit in the treatment of chronic inflammatory conditions. However, TNFalpha plays an important role in host defenses against infection, and blocking TNFalpha production may also have adverse effects. We tested the efficacy and safety of anti-TNFalpha therapy in experimental colitis induced by trinitrobenzenesulfonic acid. We cultured colonic wall specimens for bacterial growth and measured native TNFalpha protein synthesis in colonic tissue at days 0, 1, 4, 10 and 18 after induction of colitis. Anti-TNFalpha therapy (monoclonal g1 immunoglobulin, 15 mg/kg i.p., every third day) was started on either day 4 or day 10 after induction of colitis. On day 18, we measured the release of inflammatory mediators and scored colonic lesions. In acute lesions, several species of the common flora were grown, including Streptococcus, Staphylococcus, Bacteroides, clostridia and enterobacteria. In chronic lesions, only enterobacteria, clostridia and lactobacilli were isolated. TNFalpha production by inflamed colonic tissue was increased in both acute and chronic lesions. Anti-TNFalpha therapy induced a significant decrease in the release of inflammatory mediators and histopathological remission when treatment started on day 10. However, anti-TNFalpha therapy increased eicosanoid release and lesion scores when treatment started on day 4. In conclusion, acute colonic lesions showed polymicrobial infection. Anti-TNFalpha therapy induced remission of chronic intestinal inflammation, but early treatment did not prove effective.