Impact of Probiotics BIFILACT® on Diarrhea in Patients Treated With Pelvic Radiation

This study was prospective, unicentric, controlled with comparison to a placebo, randomised and double-blinded. Patients were stratified into three groups: prostatic cancers, gynecologic cancers without chemotherapy and gynecologic or rectal cancers with chemotherapy. Patients were block-randomized by the research nurse and received a standard dose of Bifilact® probiotics or placebo according to the random list generated by blocks of 2, 4, or 6 patients according to random permutations. A preliminary internal pilot study was performed when the half of the expected sample size were recruited. The goal of this analysis was to test assumptions that lead to chosen sample sizes (N0) and thus, validate the feasibility of the study. Following preliminary results, the sample size re-calculated at the interim analysis is similar to that originally intended. It would be advisable to continue the recruitment as planned. No adjustment to the critical level of the test would be required in analyzes carried out at the end of study. However, another random block using higher probiotics dosage to the randomization was added with preservation of double bind. New random list were generated for each strata with a 3:1:1 ratio (higher dose, standard dose, placebo) to compensate for the starting of the latest higher dose group. All the bottles had a similar aspect, they were all identified by the commercial brand Bifilact. Also the group, either A, B or C, was circled on that bottle, depending on if that bottle belong to the placebo group, standard dose group, or high dose group. Only the nurse knew the coding system, it was also her who would assign the patient to a group, according to the randomization list. She would also hand out the bottle to the patient every week. The two registered dietician, caregivers, were totally excluded of this processes (preliminary study and randomisation) to preserve the double blind.