A review of dose-responses of probiotics in human studies.

The probiotic definition requires the administration of an 'adequate amount' in order to obtain a health benefit. What that amount should be is not indicated. Here, an overview is given of studies that investigated the dose-response relation of probiotics in human interventions. Studies were divided in; meta-analyses, meta-analyses on specific probiotic strains, and studies testing two or more doses of a probiotic (combination) in the same study. Meta-analyses on the effect of probiotics on antibiotic associated diarrhoea (AAD) suggest a dose-response effect; for Clostridium difficile-associated diarrhoea on the other hand no dose-response was observed. For other end-points; such as necrotising enterocolitis, prevention of atopic dermatitis and slow intestinal transit, no dose-response relation was identified in meta-analyses. For prophylaxis in colorectal cancer and relief of irritable bowel syndrome, no dose-response relation was determined. However, for blood pressure, a meta-analysis observed that higher doses (greater than 1011 cfu) were more effective than lower doses. Meta-analyses of specific strains suggest a break-point for effectiveness of Lactobacillus rhamnosus GG in the treatment of acute gastroenteritis in children; no dose-response was observed for two other probiotics assessed. Studies comparing two or more doses indicate that faecal recovery and risk reduction of AAD follow a positive dose-response relationship. Other end-points such as immune markers, general health, and bowel function did not exhibit clear dose-response relations. For AAD, the findings are very compelling; both meta-analyses and dedicated dose-response studies observe a positive correlation between dose and AAD risk. These findings do not allow for extrapolation, but suggest that studying higher doses for this end-point would be worthwhile. The lack of a clear dose-response for other end-points, does not mean it does not exist; present data does just not allow drawing any conclusions.