that in order for something to be considered a probiotic (or prebiotic), its health effects and safety must be demonstrated, with an emphasis on data from human studies, and strains must be clearly identified and deposited in public culture collections. Because of the uniqueness of each probiotic (and prebiotic) strain, the health effects for each individual strain or strain combination should be documented separately, according to Salminen.

Yet a number of bacteria currently being marketed in the European Union as probiotics have no demonstrated health-promoting properties, and different strain combinations are advertised without any proven association with health benefits. One of the goals of the 2006 health claim regulation was to improve consumer protection by more clearly identifying actual probiotic and prebiotic products and their benefits to the consumer.

As part of its task, EFSA was required to assess existing health claims in each individual EU member state. These assessments required identifying and characterizing the probiotics in use; evaluating relevant studies, with an emphasis on controlled human intervention studies; and assessing whether the proposed health relationship is something that consumers can understand.

Characterization alone has been a challenge. The purpose of characterization is to assure EFSA that the substance for which a claim is made is the same as that for which the evidence on efficacy is provided. Until a substance is characterized, EFSA cannot conduct a health claim assessment. Deposit to an international culture collection is key. The strain does not have to be publicly available, Salminen noted, but it should be available to regulatory officials. In its assessment of existing health claims, EFSA has identified more than 100 probiotic products that could not be characterized because of a lack of data on the strain used.

The most important component of a health claim assessment is human intervention studies (van Loveren et al., 2012). For disease risk reduction claims, the manufacturer needs to show that the product causes change in a generally accepted risk factor in a normal, healthy population, Salminen noted. It is also important for human intervention studies to be supported by animal model or other mechanistic studies.

Salminen acknowledged the challenge of demonstrating a change in a normal, healthy population. To illustrate the difficulty, he mentioned a study on a milk-based drink containing a combination of L. rhamnosus CG, L. rhamnosus Lc705, Propionibacterium freudenreichii ssp. shermannii JS, and a Bifidobacterium strain. The goal was to see if a daily dose of 4 × 109 CFU, with equal amounts of each bacterial strain, would reduce GI discomfort in a normal, healthy population of individuals. According to Salminen, although two “quite nice” clinical studies were conducted (both randomized, placebo-controlled, double-blind intervention studies), they were conducted using two different strains of Bifidobacterium (EFSA,



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement