How Strong Should a Probiotic Be? CFU, Billions and What Trials Used

There is no single right strength. The effective doses in the trials ran from 100 million to 100 billion CFU a day depending on the strain and the outcome, and the only outcome with a clear dose response is diarrhoea during a course of antibiotics, where children given 5 billion CFU a day or more did better than those given less.^{1, 2} This page is one spoke of the Wellgard probiotics guide.

Evidence at a glance

• Trials: nine trials and pooled analyses with a stated dose, from a four-week IBS trial of 362 women at 100 million CFU a day to a three-week tolerance trial at 100 billion.^{3, 4}
• Largest single trial: 2,981 UK inpatients aged 65 and over, 60 billion organisms a day for 21 days, no difference from placebo.⁵
• Strongest measured outcome: in children's antibiotic-course trials, diarrhoea in 8% on 5 billion CFU a day or more against 23% on placebo (relative risk 0.37), compared with 0.68 at lower doses.²
• Not yet tested: a count above which side effects rise; no trial has found one, and none has compared two shop products head to head.

Sources checked on 16 October 2026.

How strong should a probiotic be?

As strong as the trial for your outcome. "Strength" on a label is the CFU, colony-forming units: the number of cells that can still grow when the product is tested. Across the trials behind this blog the effective counts run from 100 million a day in one IBS trial to 100 billion in the one tolerance trial that tested the ceiling.^{3, 4} The old rule of "at least 1 billion, 10 to 20 billion for adults" appears in no trial.

How many billion probiotics should I take?

The table gives the doses the trials used, by outcome.

A 2017 review of dose-response studies, by an author employed by a culture supplier, found the same pattern: a clear dose effect for antibiotic-associated diarrhoea and none established for most other outcomes.¹

What does the number on the label actually tell you?

Less than it looks. Three points about the category and the law.

• UK rules do not say whether the count is at manufacture or at best-before; ask which. Dried cultures lose cells to moisture, oxygen and heat: in a 14-month laboratory study of one strain at room temperature, the loss ran from about half the cells at low moisture to over 99% at slightly higher moisture.⁸ A count guaranteed to the best-before date is the figure that tells you what you will take.
• A total count says nothing about the split between species; the law does not require it to be printed.
• A strain code names the exact culture that was studied; a species name tells you the family. Species-only labelling is lawful.

Reading a label against the trials is on which probiotic is right for your needs.

How many strains should a probiotic have?

The number of strains is not a measure of quality. A 2021 review of 65 trials in 10,863 people found that where a single strain and a mixture containing it had been compared for the same condition, they were usually equivalent, with one result each way.⁹ The question is whether the strains on the label, at the amount on the label, have a trial for the thing you want.

Can probiotics be too strong?

No trial has found a count at which side effects rise. At 100 billion CFU a day for three weeks, 34 healthy adults recorded no change in gut symptoms, stool habit or liver and kidney blood tests; the trial was run by a culture manufacturer.⁴ Across 36 adult antibiotic-course trials in 9,312 people, adverse events on live cultures matched placebo (relative risk 1.00) in every dose subgroup.¹⁰ The groups who ask the clinical team before any live culture are the same at every count: a weakened immune system, serious illness, a central line in hospital.¹¹

How do you know if a probiotic is too strong for you?

You cannot tell from symptoms: wind, bloating and a looser stool in the first days are recorded at every dose and at the same rate as placebo.^{10, 12} In the only IBS trial to compare three doses, the high dose showed no benefit because the capsule had formulation problems, not because 10 billion was too much.³ The useful test is time: a capsule still causing symptoms after two weeks is one to stop. What the first fortnight looks like is on feeling worse before better.

Is 50 billion or 100 billion CFU too much?

Not in the trial record. 100 billion a day was tolerated for three weeks with no change in the bowel diary,⁴ and 60 billion organisms a day for 21 days in nearly 3,000 UK hospital patients aged 65 and over produced no serious adverse event attributed to it, and no difference from placebo in diarrhoea.⁵ Nor is a higher count automatically better: the same result at 1 and 10 billion for constipation, and 100 million beating 10 billion for IBS.^{3, 6} A looser stool at a high count is covered on can probiotics cause diarrhoea.

What we don't know

• No trial has found a ceiling; the highest daily count tested in healthy adults is 100 billion, for three weeks.⁴
• Outside antibiotic-associated diarrhoea, dose comparisons are rare and give the same result at low and high counts.¹
• No study has compared blister packs with jars for how many cells survive.

Frequently asked questions

Is 1 billion CFU enough?

For some outcomes, close to it: 1.8 billion a day shortened transit in the 14-day trial, and across 52 IBS trials benefit appeared from 1 billion a day, from four weeks.^{13, 7} For children's antibiotic diarrhoea, 5 billion a day or more did better.²

Is 3 billion CFU enough?

It depends on the outcome: the trials that measured an effect used a stated count of a named strain for a fixed period, and none compared 3 billion with a higher count for the same outcome. The counts behind this blog run from 100 million a day in the four-week IBS trial of 362 women, where that dose improved symptoms and the 10 billion dose did not, to 10 billion a day in the one-year bone trial in older women (a two-year trial in 239 younger women then found no difference) and 100 billion in the tolerance trial; for children's antibiotic-associated diarrhoea, 5 billion a day or more did better than lower doses. Matching the label count to a trial for your outcome is the test, not the size of the number.^{3, 14, 15, 4, 2}

What is the maximum amount of probiotics per day?

No authority sets one and no trial has found one; the highest daily count tested in healthy adults is 100 billion CFU for three weeks, without effect on the bowel.⁴

Sources

1. Ouwehand AC. A review of dose-responses of probiotics in human studies. Beneficial Microbes, 2017. Review; author employed by a culture supplier. PubMed 28008787.
2. Guo Q, et al. Probiotics for the prevention of pediatric antibiotic-associated diarrhea. Cochrane Database of Systematic Reviews, 2019. Pooled analysis of 33 randomised trials, 6,352 children. PubMed 31039287.
3. Whorwell PJ, et al. The American Journal of Gastroenterology, 2006 (an encapsulated bifidobacterium in women with irritable bowel syndrome). Randomised trial, 362 women, 4 weeks; funding not stated in the abstract. PubMed 16863564.
4. Wind RD, et al. British Journal of Nutrition, 2010 (tolerance and safety of a Lactobacillus rhamnosus strain in healthy volunteers). Randomised double-blind trial, 34 healthy adults, 3 weeks; run by a culture manufacturer. PubMed 20691131.
5. Allen SJ, et al. Lactobacilli and bifidobacteria in the prevention of antibiotic-associated diarrhoea and Clostridium difficile diarrhoea in older inpatients (PLACIDE). The Lancet, 2013. Randomised double-blind trial, 2,981 UK inpatients aged 65 and over, 21 days; publicly funded. PubMed 23932219.
6. Ibarra A, et al. Gut Microbes, 2018 (28-day supplementation with a Bifidobacterium lactis strain in adults with functional constipation). Randomised double-blind dose-ranging trial, 228 adults, 28 days; run by the strain's manufacturer. PubMed 29227175.
7. Waller PA, et al. Scandinavian Journal of Gastroenterology, 2011 (dose-response of a Bifidobacterium lactis strain on whole-gut transit time). Randomised trial, 100 adults, 14 days; strain supplied by its manufacturer, funding not checked. PubMed 21663486.
8. Vesterlund S, et al. International Journal of Food Microbiology, 2012 (water activity and the storage stability of a Lactobacillus rhamnosus strain in a dry food matrix). Laboratory storage study, 14 months. PubMed 22663981.
9. McFarland LV. Efficacy of Single-Strain Probiotics Versus Multi-Strain Mixtures: Systematic Review of Strain and Disease Specificity. Digestive Diseases and Sciences, 2021. Systematic review of 65 randomised trials, 10,863 people. PubMed 32274669.
10. Liao W, et al. Probiotics for the Prevention of Antibiotic-associated Diarrhea in Adults: A Meta-Analysis of Randomized Placebo-Controlled Trials. Journal of Clinical Gastroenterology, 2021. Pooled analysis of 36 randomised trials, 9,312 adults. PubMed 33234881.
11. Doron S, Snydman DR. Risk and safety of probiotics. Clinical Infectious Diseases, 2015. Review. PubMed 25922398.
12. Esmaeilinezhad Z, et al. Probiotics for the prevention of Clostridioides difficile-associated diarrhea in adults and children. Cochrane Database of Systematic Reviews, 2025. Cochrane review; adverse events from 37 randomised trials, 11,911 people. PubMed 40931979.
13. Zhang WX, et al. Efficacy of probiotics, prebiotics and synbiotics in irritable bowel syndrome: a systematic review and meta-analysis of randomized, double-blind, placebo-controlled trials. Journal of Medical Microbiology, 2023. Pooled analysis of 52 randomised trials, 6,289 people. PubMed 37772692.
14. Nilsson AG, et al. Lactobacillus reuteri reduces bone loss in older women with low bone mineral density: a randomized, placebo-controlled, double-blind, clinical trial. Journal of Internal Medicine, 2018. Randomised trial, 90 women aged 75 to 80, 12 months; funding not checked. PubMed 29926979.
15. Gregori G, et al. JAMA Network Open, 2024 (a Limosilactobacillus reuteri strain and early postmenopausal bone loss; randomised clinical trial). Randomised double-blind trial, 239 women aged 50 to 60, 2 years. PubMed 38865129.

Spotted something out of date? Email support@wellgard.co.uk with the page address and the sentence.