The probiotics evidence splits by indication: null for gastroenteritis, positive for colds

"Do probiotics do anything" has no single answer, because the trials answer different questions and they disagree with each other. For acute infectious diarrhoea — the use with the longest history and the strongest marketing — two large randomised trials and a Cochrane review of 82 studies found essentially nothing. For preventing upper respiratory infections, a separate Cochrane review of 23 trials found a real reduction, on evidence the reviewers themselves graded low-certainty.

The category is not one intervention. It is a shelf of different organisms tested against different illnesses, and the results do not travel between them.

The gastroenteritis result, twice

In 2018 two independent trials of the same question were published in the same issue of the New England Journal of Medicine. Both were randomised, double-blind and placebo-controlled, and both recruited from paediatric emergency departments.

The American trial enrolled 971 children aged 3 months to 4 years with acute gastroenteritis across 10 US paediatric emergency departments, giving a five-day course of Lactobacillus rhamnosus GG at 1×10¹⁰ colony-forming units twice daily or matching placebo [s1]. The primary outcome — a modified Vesikari scale score of 9 or higher within 14 days, indicating moderate-to-severe illness — occurred in 55 of 468 children on the probiotic (11.8%) and 60 of 475 on placebo (12.6%), a relative risk of 0.96 (95% CI 0.68 to 1.35, P=0.83) [s1]. Median diarrhoea duration was 49.7 hours with the probiotic and 50.9 hours with placebo [s1]. Vomiting duration, day-care absence and household transmission also showed no significant differences [s1].

The Canadian trial enrolled 886 children aged 3 to 48 months across six paediatric emergency departments, testing a five-day combination of Lactobacillus rhamnosus R0011 and L. helveticus R0052 at 4.0×10⁹ colony-forming units twice daily [s2]. Moderate-to-severe gastroenteritis within 14 days occurred in 108 of 414 on the probiotic (26.1%) and 102 of 413 on placebo (24.7%), an odds ratio of 1.06 (95% CI 0.77 to 1.46, P=0.72) [s2]. Median diarrhoea duration was 52.5 hours against 55.5 hours [s2].

Two trials, two continents, two different products, same answer.

What the pooled evidence adds

A 2020 Cochrane review pooled 82 studies with 12,127 participants — 11,526 of them children — on probiotics for acute infectious diarrhoea [s3]. Its headline caution is methodological: risk of bias was high or unclear in many studies, statistical heterogeneity was marked, and funnel plots demonstrated publication bias for the main outcomes [s3].

The reviewers therefore ran their principal analysis restricted to studies at low risk of bias across all indices. There, probiotics made no detectable difference to the risk of diarrhoea lasting 48 hours or longer (risk ratio 1.00, 95% CI 0.91 to 1.09; 2 trials, 1,770 participants; moderate-certainty evidence), and the effect on duration was too uncertain to call — a mean difference of 8.64 hours shorter, with a confidence interval running from 29.4 hours shorter to 12.1 hours longer, graded very low certainty [s3]. Restricting to the two most-studied organisms, Lactobacillus rhamnosus GG and Saccharomyces boulardii, did not change the picture [s3]. The reviewers' conclusion is that probiotics probably make little or no difference to how many people have diarrhoea lasting 48 hours or more [s3]. No serious adverse events were attributed to probiotics [s3].

Where the signal is

A different Cochrane review, updated in 2022, asked whether probiotics prevent acute upper respiratory tract infections, pooling 23 trials with 6,950 participants spanning children from one month old, adults with a mean age of 37.3 years, and older people with a mean age of 84.6 [s4]. Most trials used one or two strains for more than three months [s4].

Here the results moved. Probiotics may reduce the number of participants diagnosed with at least one upper respiratory infection (risk ratio 0.76, 95% CI 0.67 to 0.87; 16 studies, 4,798 participants; low-certainty evidence), and likely reduce the number having at least three episodes (risk ratio 0.59, 95% CI 0.38 to 0.91; 4 studies; moderate certainty) [s4]. The incidence rate fell (rate ratio 0.82, 95% CI 0.73 to 0.92), episodes were shorter by a mean 1.22 days (95% CI 0.33 to 2.12 days shorter), and — the outcome with the widest public-health relevance — fewer participants used prescribed antibiotics for these infections (risk ratio 0.58, 95% CI 0.42 to 0.81; 6 studies; moderate certainty) [s4]. Adverse events were minor and no more common than with placebo (risk ratio 1.02, 95% CI 0.90 to 1.15) [s4].

The reviewers' summary is that probiotics were better than placebo or no treatment at preventing acute upper respiratory infections [s4]. Most of those estimates carry a low-certainty grade, which in GRADE terms means further research is likely to change them.

How to hold the two results together

The honest reading is that "probiotics" is not a thing that works or does not work. A specific organism, at a specific dose, given for a specific purpose, either has trial evidence behind it or does not — and for the single best-tested consumer use, treating a child's stomach bug, two large well-conducted trials and a low-risk-of-bias meta-analysis all landed on no benefit [s1] [s2] [s3]. For a different purpose, in trials of different strains taken continuously for months rather than days, the pooled estimate favours probiotics but the certainty is low [s4].

Product labels rarely make that distinction, and strains studied in trials are not always the strains sold. Whether any particular product has evidence behind it for any particular purpose is a question to put to a clinician or pharmacist rather than to a category.