Objective: Probiotics are live bacteria or non-pathogenic yeasts that colonize the gastrointestinal tract, providing health benefits to the host. Their safety has not been systematically investigated in pregnancy. Our objective was to review the evidence for safety of Lactobacillus, Bifidobacterium, and Saccharomyces during pregnancy. Methods: Accepted were randomized controlled trials comparing probiotics and placebo given any time during pregnancy, at any dose, in any form, for any condition, and given =1 week. Outcomes included birth weight, gestational age(GA), rates of C-section, malformations (major/minor), and miscarriage. Two independent reviewers searched databases (MEDLINE, OLDMEDLINE, CINAHL, Cochrane Database, DARE, Allied and Complementary Medicine, EMBASE, AltHealthWatch, American Botanical Council, Natural Database, and Natural Standard) from inception to September 2007, hand-searched their bibliographies, and contacted experts to identify other trials. Two reviewers extracted data and assessed study quality using the Downs-Black checklist. Discrepancies (study eligibility, data, quality) were adjudicated through consensus. Heterogeneity was assessed with chi-squared and I-squared tests, publication bias with Begg-Mazumdar test. Random-effects models combined data. Results: Nine studies examined 1706 patients for three outcomes; none examined birth defects or miscarriages. Publication bias was not detected (tau=0.40,p=0.33). Quality averaged 62%±20%. Five studies (648 exposed,641 controls) reporting C-section outcomes found no heterogeneity (chi2=1.43,P=0.84;I2=0). The odds ratio was 0.90 (95%CI:0.67-1.22). Birth weights were examined from 1718 patients (851 exposed,867 controls), no heterogeneity was found (chi2=0.16,P=1.0;I2=0). Exposed infants averaged 3718 (CI95%:3529-3906) grams, controls weighed 3620 (CI95%:3423-3817) grams, all within normal range. The weighted difference was 83 (CI95%:-109 to 275) grams. Four studies examined GA in 347 patients (no heterogeneity; chi2=1.31,P=0.73;I2=0). The mean GA of exposed infants was 39.6 (CI95%:39.2-40.0) weeks, controls were 39.4 (CI95%:39.0-39.8) weeks, all within the normal range. The weighted difference was 0.2 (CI95%:-0.2 to 0.6) weeks. Conclusion: Administering Lactobacillus and Bifidobacterium had no effect of C-section incidence, birth weight or GA.