Probiotics for the prevention of antibiotic-associated diarrhea: an umbrella review and second-order meta-analysis of 21 systematic reviews covering 170 unique randomized controlled trials

Background Antibiotic-associated diarrhea (AAD) is a common adverse event of antimicrobial therapy. Probiotics have been investigated for the prevention of AAD, but the evidence remains heterogeneous and fragmented. We conducted an umbrella review and second-order meta-analysis to comprehensively evaluate the efficacy and safety of probiotics for AAD prevention. Methods We searched PubMed, Embase, Web of Science, the Cochrane Library, the Cochrane Database of Systematic Reviews, CINAHL, CENTRAL, and Scopus from inception to March 21, 2026. Systematic reviews (SRs) with meta analyses (MAs) of randomized controlled trials (RCTs) evaluating oral probiotics for AAD prevention were included. Two reviewers independently screened studies, extracted data, and assessed methodological quality using AMSTAR 2. Overlap of primary studies was quantified with the corrected covered area (CCA) and the GROOVE tool. A second-order meta-analysis of review-level effect estimates was performed using random-effects models, with sensitivity analyses addressing review quality and primary-study overlap. Reporting followed PRIOR and PRISMA 2020. Results Twenty-one SRs/MAs were included; together they contained 478 RCTs, corresponding to 170 unique RCTs. Most SRs/MAs were rated low or critically low quality by AMSTAR 2. The CCA indicated moderate overall overlap (9.06%), but 71 of 210 pairs of reviews showed very high overlap (≥15%). In the pooled analysis, probiotics were associated with a lower risk of AAD (RR 0.54, 95% CI 0.48–0.61; OR 0.37, 95% CI 0.31–0.46). Because the SRs/MAs share many primary trials, these estimates are probably more precise than the evidence warrants. Estimates were similar in leave-one-review-out analyses (RR 0.54–0.55) and when restricted to SRs/MAs rated high or low quality (RR 0.52, 95% CI 0.48–0.57), but weaker and less precise in subsets of SRs/MAs with low pairwise overlap (RR 0.59–0.63). Pooled estimates were similar in children (RR 0.46) and adults (RR 0.55), whereas in the four SRs/MAs reporting on older adults no clear benefit was seen (RR 0.96, 95% CI 0.85–1.08). In an exploratory comparison based on three overlapping SRs/MAs, the estimate was lower for daily doses ≥5 × 10 9 CFU than for lower doses; this comparison is confounded by age and product and cannot be applied to individual products. No increase in adverse events was detected, but only two SRs/MAs reported this outcome. Conclusion Across published SRs/MAs, probiotics were associated with a lower risk of antibiotic-associated diarrhea, without evidence of increased adverse events. However, only 1 of 21 SRs/MAs was rated high quality and 13 were rated critically low, the SRs/MAs share many primary trials, and the interventions were heterogeneous in strain, dose, and formulation. The findings therefore represent low-certainty evidence, do not show that any specific probiotic product is effective, and should not be used to support class-wide or product-specific claims. High-quality, strain-specific RCTs and well-conducted, registered SRs/MAs are needed. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/view/CRD420261404660 , identifier PROSPERO (CRD420261404660).