Probiotics in children: what the research shows

"Do probiotics work for children?" sounds like a simple question. The research literature suggests it is the wrong one. A probiotic is not a single intervention. Evidence attaches to a specific strain, given at a specific dose, for a specific condition, in a specific group of children, and a result for one combination does not automatically carry over to another.

This article walks through two bodies of pediatric evidence that point in different directions, and uses them to explain how researchers read trials like these. It is educational and not a substitute for personal medical advice. Decisions about any child's care belong with that child's own clinician.

What does research show about probiotics in children?

The short answer is that it depends on the strain, the dose, and the condition. Pooled trials suggest that certain probiotics moderately reduce antibiotic-associated diarrhea in children. Two large, rigorous trials of probiotics for acute gastroenteritis, published together in 2018, found no benefit over placebo. Both findings are real. They answer different questions about different products.

Why is "do probiotics work?" the wrong question?

Start with a definition. A probiotic is a live microorganism that, when given in adequate amounts, confers a health benefit on the host. That is the definition used in the 2023 position paper of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN), discussed below (Szajewska et al., 2023).

Notice what the definition does not say. It does not say that every live microorganism sold as a probiotic confers a benefit. It says the benefit is part of what earns the name.

The second idea is strain specificity. The National Institutes of Health's National Center for Complementary and Integrative Health (NCCIH) puts it plainly: if a specific kind of Lactobacillus helps prevent an illness, that does not necessarily mean another kind of Lactobacillus, or any Bifidobacterium, would do the same (NIH NCCIH). A trial result belongs to the product that was tested.

So a well-formed research question has four parts: which strain, at what dose, for which condition, in which children. Change any one of them and the evidence may change with it.

What did the largest trials in acute gastroenteritis find?

Acute gastroenteritis is a short-term infection or inflammation of the stomach and intestines, usually causing diarrhea and vomiting. In November 2018, the New England Journal of Medicine published two large trials of probiotics for young children with this illness, both run in pediatric emergency departments.

The first, by Schnadower and colleagues (2018), enrolled 971 children aged 3 months to 4 years at 10 US pediatric emergency departments. Children received a 5-day course of Lactobacillus rhamnosus GG or a matching placebo. Moderate-to-severe gastroenteritis within 14 days occurred in 11.8% of the probiotic group and 12.6% of the placebo group (relative risk 0.96, 95% confidence interval 0.68 to 1.35). There were no significant differences in the duration of diarrhea or vomiting, in day-care absence, or in household transmission.

The second, by Freedman and colleagues (2018), enrolled 886 children aged 3 to 48 months at six Canadian pediatric emergency departments. Children received a combination product containing L. rhamnosus R0011 and L. helveticus R0052, or placebo. Moderate-to-severe gastroenteritis occurred in 26.1% of the probiotic group and 24.7% of the placebo group (odds ratio 1.06, 95% CI 0.77 to 1.46).

A few terms explain why these trials carry weight:

• Randomized means chance, not a clinician or parent, decided which child received the probiotic, so the groups should be comparable in ways no one measured.
• Double-blind means neither the families nor the study staff knew which child got which product, which protects the outcome from expectation.
• Placebo-controlled means the comparison group received an identical-looking product with no active organisms.
• A prespecified primary outcome is the single main result the investigators committed to before seeing the data. In both trials it was moderate-to-severe illness defined by a score of 9 or higher on the modified Vesikari scale, a standardized severity score that ranges from 0 to 20.

Two large, well-blinded trials with a committed outcome and a null result carry real weight against earlier, smaller trials that suggested benefit. That does not prove every earlier study wrong, or that no strain could ever help in any setting. It does show that the two products tested, in these children and settings, did not outperform placebo on the outcome that mattered most. The site's article on why replication matters in science covers this pattern in more depth.

Why does the evidence look different for antibiotic-associated diarrhea?

Antibiotic-associated diarrhea is diarrhea that develops during or after a course of antibiotics, which can disturb the normal balance of gut microbes. Here the question is prevention during antibiotic treatment, not treatment of an infection already under way.

A Cochrane systematic review by Guo and colleagues (2019) pooled 33 randomized trials including 6,352 children. Antibiotic-associated diarrhea occurred in 8% of children given probiotics and 19% of controls (risk ratio 0.45, 95% CI 0.36 to 0.56). The authors rated this moderate-certainty evidence of a moderate protective effect.

The review also reports a number needed to treat of 9 (95% CI 7 to 13). In plain terms, across these trials, about nine children needed to receive a probiotic for one additional child to avoid antibiotic-associated diarrhea. The site's explainer on relative versus absolute risk shows why this kind of number is often more informative than a ratio alone.

The review is worth reading as critically as its authors read it:

• Risk of bias. Twenty of the 33 trials were judged at high risk of bias, meaning their design or reporting left room for systematic error.
• Heterogeneity. The trials differed from one another more than chance alone would explain (an I² of 57%), which is expected when many different organisms, doses, and antibiotics are pooled.
• Loss to follow-up. In 19 trials, between 1% and 46% of participants were not followed to the end. When the authors made conservative assumptions about those children, the benefit remained statistically significant but smaller.
• Dose. A subgroup analysis the authors judged credible suggested higher doses were more effective than lower ones.

Their own conclusion is measured: the benefit of high-dose probiotics needs to be confirmed in a large, well-designed multicenter randomized trial, and it is premature to draw firm conclusions about other probiotic agents.

How do expert groups handle this much uncertainty?

Professional societies have to turn mixed evidence into recommendations, and they have not always agreed. The ESPGHAN paper acknowledges this directly, describing probiotics as widely used despite uncertainty about their efficacy and "discordant recommendations" about their use (Szajewska et al., 2023).

The ESPGHAN group's method is instructive. It reviewed systematic reviews and randomized trials published through December 2021, and it made a recommendation only where at least two randomized trials existed on a similar, well-defined strain. Consensus was reached through a modified Delphi process, a structured method in which experts rate statements in repeated rounds until agreement is reached.

The group did make strain-specific recommendations for several pediatric gastrointestinal conditions. Its conclusion, however, keeps the uncertainty in view: further studies confirming effects and defining the type, dose, and timing of probiotics are still often required, and the use of probiotics with no documented health benefit should be discouraged. This article deliberately does not list which strains were recommended for which condition. That is the domain of a child's clinician, not a general explainer.

What is known about safety, and where are the gaps?

In the Cochrane review, none of the 24 trials that reported on adverse events (4,415 participants) found a serious adverse event attributable to probiotics. Minor events such as rash, nausea, gas, bloating, and constipation were reported, and the certainty of the adverse event evidence was rated low (Guo et al., 2019).

That reassurance has limits. The same review notes that observational studies outside its scope have reported serious adverse events in severely debilitated or immunocompromised children with underlying risk factors. NCCIH also reports that in 2023 the US Food and Drug Administration warned health care providers that premature infants given probiotics are at risk of severe, potentially fatal infections caused by the organisms in the products (NIH NCCIH).

The general lesson is that safety observed in generally healthy trial populations does not automatically transfer to more vulnerable groups. The site's article on how safety trials in children are designed explains why.

Regulation adds one more gap between research and practice. NCCIH notes that many probiotics are sold as dietary supplements, which do not require FDA approval before they are marketed. A product on a shelf is therefore not necessarily the strain, dose, or formulation that was tested in a trial. This site's broader pediatric health and GI research program sits in the same territory, where careful study design matters as much as the headline result.

How should a careful reader weigh a new probiotic study?

Five questions do most of the work:

1. Which strain, and at what dose? A result for one strain is not a result for "probiotics."
2. Which condition, and which children? Prevention during antibiotics is a different question from treatment of an infection, and healthy children differ from vulnerable ones.
3. How large, and how well blinded? Small or open-label studies are more easily moved by chance and expectation.
4. Was the primary outcome prespecified? A benefit found only in a secondary outcome is weaker evidence.
5. Who funded it, and has it been replicated? Independent confirmation in a second well-designed trial is the strongest test.

Probiotics in children is a field where the honest summary is "it depends," followed by a clear account of what it depends on. Readers interested in how pediatric research questions are framed and tested can explore the peer-reviewed publications listed on this site.