Probiotics and the irritable bowel syndrome

Sirs, We read with interest the article by Kajander et al.1 reporting the effect of a multispecies probiotic on irritable bowel syndrome (IBS) symptoms, intestinal microbiota stability and health-related quality of life (HRQOL). We wish to raise several points warranting discussion. First, the questionnaire used for assessing HRQOL2 has not been, to our knowledge, fully validated according to current standards, especially in terms of construct validity, reliability or responsiveness.3 Secondly, patients were not comparable at baseline for the primary and several secondary end points: the composite and each of the four IBS symptom scores were systematically more severe in the probiotic than in the placebo group. Symptom severity at baseline is a key point as it influences the improvement in symptoms; the most severe IBS patients may show the largest reduction in IBS symptom severity score in part because of the regression to the mean.4, 5 Thirdly, Kajander et al.1 did not disclose or compare the proportion of responders in each group, an analysis recommended together with that of the mean change in score, by Rome III experts6 and regulatory authorities.7 Fourthly, when discussing published probiotic trials in IBS, which assessed HRQOL, Kajander et al.1 omitted to quote at this point the trial, in other respects cited in the Reference section, recently published in this Journal,8 reporting the beneficial effect halfway through the study of a fermented milk containing the Bifidobacterium animalis DN-173 010 strain on the responder rate for HRQOL abdominal discomfort dimension score: indeed, this study8 was the first randomized trial on IBS using a HRQOL domain as a primary end point. Finally, the conclusion by Kajander et al.,1 in their Discussion section, that the test probiotic ‘is effective in… improving HRQOL’ may be misleading, as (i) the discomfort score showed, in a 1–7 scale, a 0.25 point improvement only and (ii) this discomfort score is mostly symptom-based while the three other domains of the HRQOL questionnaire did not show differences between groups. The use of ‘discomfort HRQOL domain’ should have made the conclusion on HRQOL more reliable. Declaration of personal interests: O. Chassany and C. Matuchansky have served as advisory board members for Danone Research for the study referenced No. 8.