The aim of the present study was to examine if oral administration of Gly-X-Y repeat sequences alleviates disease activity in rheumatoid arthritis (RA). The study had a randomized, placebo controlled and double blind design with a wash-out/cross-over between the two 3-months long treatment periods. A total of 40 patients entered and 36 patients fulfilled the study, among them 16 started with the active drug and 20 with a placebo. Disease activity score (DAS) was used as the primary outcome measure with several secondary outcome variables. Type I or alpha error of 0.05 was accepted and the power (= 1-beta) was set to 80%, which according to the power analysis was also achieved. With active drug treatment, joint swelling score (54 count; P < 0.001), Ritchie's index (P < 0.01) and DAS (P < 0.001) improved. HAQ also improved (P < 0.05), but there was no improvement in the subjective condition of the patients as measured with the self-reported Pain Disability Index and Comprehensible Psychopathological Rating scale questionnaires. Apparently, 5/36 patients had a response of > or = 1.2 in DAS and 33/36 changed for the better; DAS impaired only in 3/36 patients during the active drug treatment When the stringent EU response criteria were applied and the results were compared to the placebo group, the response was not clinically significant. We conclude that Gly-X-Y repeat sequences are not effective as used in the present study. However, this does not definitely disprove the value of the Gly-X-Y repeat sequences, because confounding effects of dosage, concomitant medication and excessive degradation of the linear Gly-X-Y sequences in the stomach, gut or by phagolysosomes could not be adequately controlled. The discrepancy between the favourable effects in the preliminary, open pilot study and the controlled clinical trial emphasizes the value of the DAS, EU response criteria and adequately administered controlled clinical studies.