This review focuses on the efficacy of probiotics for diarrhea in children in different settings: day-care centers, diarrhea acquired in the hospital, antibiotic-associated diarrhea, and treatment of acute infectious diarrhea. For prevention of diarrhea acquired in day-care centers, 5 randomized and placebo-controlled trials have been published. Probiotics tested were Lactobacillus GG, Bifidobacterium lactis (alone or in combination with Streptococcus thermophilus), and Lactobacillus reuteri. The evidence of their efficacy in these settings is only modest: statistically significant for some strains only and in any case of minimal to mild clinical importance. Few trials have examined the potential role of probiotics in preventing the spread of diarrhea in hospitalized children, an event most commonly due to either rotavirus or Clostridium difficile, and they have yielded conflicting results. Overall, these studies provide only weak evidence on the efficacy of probiotics. On the other hand, a large number of trials on the role of probiotics in preventing the onset of antibiotic-associated diarrhea have been published. Most commonly employed probiotics were Lactobacillus GG, Bifidobacterium spp., Streptococcus spp., and the yeast Saccharomyces boulardii. In general, these trials do show clear evidence of efficacy, with the 2 most effective strains being Lactobacillus GG and S. boulardii. Today, we have a large number of published clinical trials on the role of probiotics in treating sporadic infectious diarrhea in children, and many of them are randomized, blinded, and controlled. They consistently show a statistically significant benefit and moderate clinical benefit of a few, well-identified probiotic strains-mostly Lactobacillus GG and S. boulardii, but also L. reuteri-in the treatment of acute watery diarrhea, primarily rotaviral, in infants and young children of developed countries. Such a beneficial effect seems to result in a reduction of diarrhea duration of little more than 1 day, and to be exerted mostly on diarrhea due to rotavirus. The effect is not only strain-dependent, but also dose-dependent, with doses of at least 10 billion/d being necessary.