The following double-blind, placebo-controlled, multicenter study investigated the influence of a bacterial immunostimulant (Symbioflor 1, cells and autolysate of human Enterococcus faecalis) on the occurrence of relapses in patients with chronic recurrent bronchitis (n = 136; placebo n = 66, verum n = 70) in a 6 months treatment period and a follow-up period of 8 months, compared to placebo. Under verum 39 incidents of relapses were recorded, which was about 60% the number observed among the patients treated with placebo (66 incidents). The verum preparation exhibited superior clinical efficacy compared to placebo (p = 0.001) in the Kaplan-Meier test. This better clinical efficiency of the test preparation was particularly observed during the treatment period, with 12 vs. 27 relapses (p = 0.013), but less during the follow-up observation period, with 27 vs. 39 relapses (p = 0.127). In addition, the time span until occurrence of the first relapse was clearly longer under verum (699 days) than under placebo (334 days) and after the end of the observation period 91% of patients under verum experienced only one relapse compared to 62% in the placebo group (p = 0.01). Severity of relapses under verum was also reduced significantly (chi 2; p = 0.001. Only 4 patients under verum required antibiotic therapy compared to 13 patients under placebo. Verum was equally well tolerated as placebo, with no serious side effects in either group. No changes in laboratory tests--haematology and clinical chemistry--were observed. It can be concluded, that previously demonstrated immunomodifying effects of the test preparation have clinical relevance for the treatment of chronic recurrent bronchitis because not only the number but also the severity of acute relapses could be clearly reduced. This is discussed in view of the current literature.