Probiotics for Depression & Anxiety: Independent Evidence Review

Table of contents

• Evidence summary
• What "psychobiotics" are, and their legal status
• How it might work: the gut–brain axis
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid probiotics
• Doses studied and product quality
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Note on clustering: the strongest independent syntheses come from the UK and US. Many of the positive add-on trials come from Iran and are small. Much of the "add-on" signal comes from one London group whose studentship was part-funded by a probiotic maker. The manufacturer-linked trials involve companies based in Canada, the Netherlands, Switzerland, Japan and the UK.

What "psychobiotics" are, and their legal status

Probiotics are live microorganisms, usually bacteria and sometimes yeasts, sold as foods (yoghurt, fermented drinks) or supplements. In the US, their regulation depends on intended use: a supplement does not need FDA approval before sale, while a product sold as a drug must show safety and effectiveness in clinical trials (NCCIH). "Psychobiotic" is a marketing and research term for live organisms said to benefit mental health. It has no legal definition. The researchers who tested one such strain describe psychobiotics as microorganisms with a potential mental health benefit (Kelly 2017).

Strains, not species, are the unit that matters

A probiotic is named by genus, species and strain code, for example Lactobacillus helveticus R0052. Trials of "probiotics for depression" have tested many different strains and mixes: one company's two-strain blend (Romijn 2017), a 14-strain capsule (Nikolova 2023), an eight-strain powder (Schaub 2022), a three-strain capsule (Akkasheh 2016) and a single Bifidobacterium longum strain (Pinto-Sanchez 2017). The WFSBP/CANMAT guideline says the most suitable strains for depression "have not been confirmed" (Sarris 2022). A 2024 meta-analysis that tried to group results by strain found a benefit on one self-rated scale but not on clinician-rated scales (Rahmannia 2024). This is why we do not treat "probiotics" as one product.

Legal status by region

How it might work: the gut–brain axis

The gut and brain communicate through the vagus nerve, hormones, the immune system and chemicals made by gut microbes. The theory is that changing gut bacteria could change mood. Here is what has and has not been shown:

• Animals only: In mice, L. rhamnosus JB-1 changed GABA receptor activity in the brain and reduced stress hormones and anxiety- and depression-like behaviour. The effects disappeared when the vagus nerve was cut (Bravo 2011). In rats, L. casei Shirota blunted the stress-hormone response and stimulated vagal nerve activity (Takada 2016; Yakult-authored).
• Humans, the leap that failed: When the same JB-1 strain was given to 29 healthy men for 8 weeks, it changed nothing: mood, anxiety, stress, cortisol response, inflammation, cognition and brain-wave measures were all unaffected (Kelly 2017). The authors called their paper "Lost in translation?"
• Humans, associations: A meta-analysis of 59 case-control studies found modest differences in gut bacteria between people with psychiatric disorders and controls. There was little evidence that any pattern was specific to depression; the changes were shared across diagnoses (Nikolova 2021, JAMA Psychiatry; MRC, BBSRC and UK Department of Health grants in the PubMed record). Association does not show that the bacteria cause the depression.
• Humans, biomarkers: Small trials report changes in inflammation (CRP), insulin and BDNF with some products (Akkasheh 2016; Nikolova 2021). Others found no change in any blood marker (Romijn 2017). In the London add-on trial, microbial richness rose within the probiotic group compared with baseline but not compared with placebo; overall community make-up (beta diversity) differed between groups at week 4 but not week 8, and inflammatory markers did not differ (Nikolova 2025; first author also affiliated with ADM Protexin).
• Brain imaging: Two small trials reported changes in how the brain responds to emotional faces after a probiotic (Pinto-Sanchez 2017; Schaub 2022). These are early findings and have not been confirmed by larger studies.

Bottom line: no mechanism has been proven in humans to explain how a probiotic improves mood.

What works and what does not

Benefits by claim

Do probiotics work for depression?

The largest broad analysis pooled 34 controlled trials. Probiotics lowered depression scores by a small amount overall. Prebiotics, the fibres that feed gut bacteria, did nothing measurable (d = −0.08). The authors reported a larger effect in people with a medical or psychiatric condition, and noted that few trials studied clinical samples (Liu 2019; exact subgroup figures not confirmed in our check). In plain terms, for most healthy people any effect is too small to notice.

The most recent clinically focused review, from Oxford and funded by the UK's National Institute for Health and Care Research (NIHR), included 23 randomised trials. The abstract reports 1,401 participants, while the results section reports 1,280; we flag this discrepancy. Probiotics reduced depression scores by SMD −0.96 (95% CI −1.31 to −0.61) across 18 trials and 969 people (Asad 2025). That looks large, so the caveats matter. The trials disagreed a lot (I² = 85%), the authors reported signs of publication bias, and GRADE certainty was low.

Are probiotics useful as an add-on to antidepressants?

This is the strongest version of the claim. A King's College London team found that, across 7 trials and 404 people, probiotics added to antidepressants reduced symptoms (SMD 0.83, 0.49 to 1.17), while probiotics on their own did not (SMD −0.02) (Nikolova 2021). The trials came from Iran, Poland, Japan, New Zealand and Australia. The WFSBP/CANMAT taskforce rated adjunctive probiotics "Provisionally Recommended" for depression. It warned that the trials were small with modest effects, so probiotics could "only at present be modestly recommended" alongside standard treatment (Sarris 2022). A newer Canadian task force (CANMAT, published November 2025; no external funding and no conflicts declared) reviewed 23 trials and 8 meta-analyses. It concluded that probiotics "may be cautiously considered as third-line adjunctive treatments" for major depression, and that they "cannot currently be recommended as standalone treatments" (CANMAT 2025).

Not every review splits the evidence so neatly. In the 2025 Oxford review, stand-alone probiotics (5 trials, 266 people: SMD −0.95) and add-on probiotics (12 trials, 624 people: SMD −1.06) gave similar pooled effects, and the difference between them was not significant (Asad 2025). Both estimates carry the same low certainty, high heterogeneity and small-study concerns described above.

The same London group then ran a pilot trial. Forty-nine adults with depression who had not fully responded to an antidepressant took a multistrain ADM Protexin probiotic or placebo for 8 weeks. The authors reported greater improvements in depressive and anxiety symptoms than placebo, called the results promising and asked for a definitive trial (Nikolova 2023). The trial was funded through an MRC Industrial CASE studentship with ADM Protexin as industry partner, and the lead author reported ADM Protexin grants. The paper also states partial funding from the NIHR Maudsley Biomedical Research Centre. We did not verify the remaining disclosure details (company input and manuscript review) against the full text, and we do not quote the trial's effect sizes for that reason.

A Swiss trial with no declared commercial funding gave people with depression an eight-strain product (Vivomixx, supplied by Mendes SA) or placebo for about one month alongside usual treatment. The daily dose was 900 billion CFU. The authors reported that depression scores improved more on the probiotic among participants with high compliance, and that the effect was more evident at follow-up. They also called for a larger study (Schaub 2022). Because the benefit depended on compliance, the headline result is weaker than a simple intention-to-treat result would suggest. We could not confirm the sample size or drop-out figures from the text we accessed.

Two Iranian trials also reported benefits on the self-rated Beck Depression Inventory. In one, a three-strain capsule was given to patients, and scores fell more than on placebo (Akkasheh 2016; funding statement not accessed). In the other, 110 patients were randomised to Lallemand's two-strain product, a prebiotic or placebo; 81 finished. Beck scores fell significantly more on the probiotic than on placebo, while the prebiotic arm did not differ significantly from placebo (exact scores not confirmed in our check) (Kazemi 2019). The university funded the trial; trade press reports that Lallemand provided the products (NutraIngredients 2020).

Can probiotics replace an antidepressant?

The evidence says no, but the picture is mixed. Stand-alone use showed no benefit in clinical depression (Nikolova 2021), although the 2025 Oxford review found a stand-alone effect with low certainty (Asad 2025). The clearest single trial of stand-alone use in people with low mood is a New Zealand trial of 79 adults with at least moderate low mood who were not taking psychiatric medicines, run for 8 weeks against placebo (Romijn 2017). We could not read its full text, so we do not rely on its reported outcomes. Its reported funder is the strain maker, Lallemand.

Do probiotics help anxiety?

The broad meta-analysis found only a tiny effect on anxiety (Liu 2019). In diagnosed patients, 9 small trials with 510 people gave SMD −0.59 (−0.98 to −0.19), with high heterogeneity (Asad 2025). Most of these patients had depression or a gut disorder rather than a primary anxiety disorder. The guideline taskforce listed probiotics only under mood disorders, not anxiety disorders (Sarris 2022). In people with irritable bowel syndrome, a Nestlé strain did not change anxiety scores (Pinto-Sanchez 2017).

Do probiotics reduce stress?

The positive human stress studies are mostly from manufacturers. Yakult researchers pooled three trials in medical students. Drinking fermented milk with L. casei Shirota for 8 weeks before a national exam reduced the stress-related rise in salivary cortisol and physical symptoms (Takada 2016). A French contract laboratory reported less psychological distress in healthy volunteers after 30 days of the Lallemand strain pair (Messaoudi 2011; no funding statement in the PubMed record). Lallemand cites this study as part of the scientific basis for its product (NutraIngredients 2020). The publicly funded Irish test of a strain chosen for its strong mouse results found no effect on stress (Kelly 2017).

Can probiotics reduce negative thinking or "cognitive reactivity"?

Two trials of Winclove's multispecies product, each with a Winclove co-author, measured how strongly sad mood triggers negative thoughts. In 40 healthy volunteers over 4 weeks, cognitive reactivity fell more than on placebo (Steenbergen 2015). In 71 people with depression over 8 weeks, everyone improved, and the abstract reports a between-group difference only for cognitive reactivity, not for the depression score itself (Chahwan 2019). A change on a vulnerability questionnaire is not the same as treating depression.

Risks and all side effects

For healthy adults, probiotics in mood trials were mostly well tolerated. In the London pilot there were no serious adverse reactions. Short-lived nausea and indigestion occurred only in the probiotic group (Nikolova 2023). The WFSBP/CANMAT taskforce rated the safety and tolerability of probiotics as "robust" and "considered safe" (Sarris 2022). That reassurance comes with two large caveats.

We do not know the true rate of harms

An independent review of 384 randomised trials of probiotics, prebiotics and synbiotics found that 28% reported no harms data at all, and 90% did not define adverse events. Only 7 of 53 trials in hospitalised or critically ill patients reported the number of serious adverse events per group. The authors concluded that the evidence cannot support firm safety conclusions (Bafeta 2018; no specific funding). The mood trials are mostly 4–8 weeks long, so long-term safety for this use is untested.

Other side effects and theoretical risks

• Gut symptoms: nausea and indigestion, usually transient (Nikolova 2023).
• Antibiotic resistance genes: NCCIH lists transfer of resistance genes from probiotic organisms to other microbes as a possible risk (NCCIH).
• Harmful substances and contamination: NCCIH notes that probiotics may produce harmful substances. Some products have contained microorganisms other than those listed (NCCIH).
• Delayed treatment: the largest indirect risk is using a probiotic instead of an effective treatment for depression. Stand-alone probiotics showed no benefit in clinical depression (Nikolova 2021).

All interactions

• Antidepressants: In the add-on trials, probiotics were taken alongside antidepressants (Nikolova 2023; Schaub 2022). We found no reported drug–drug interaction, but the trials were small and short.
• Immunosuppressive treatment, chemotherapy, central lines and critical illness: these are the settings where probiotic bloodstream infections occur (NCCIH; Yelin 2019; Vinayagamoorthy 2023).
• Antibiotics and antifungals: these can in principle kill probiotic bacteria or yeasts, and blood infections from probiotic strains need appropriate antimicrobial treatment. We did not find an independent source giving timing advice, so we give none.
• Prebiotic fibres: often sold in combination ("synbiotics"). Prebiotics alone did not improve mood (Liu 2019).

Who should avoid probiotics

• Premature babies, except under a neonatal team's direction. No probiotic is FDA-approved as a drug for infants (FDA 2023).
• People with weakened immunity or serious illness, including hospital and intensive-care patients, unless a clinician advises it (NCCIH).
• Anyone using a probiotic instead of assessment for moderate or severe depression, suicidal thoughts or an anxiety disorder.
• Pregnancy, breastfeeding and children: we found no mood trials in these groups among the sources reviewed, so there is no evidence of benefit for mood. Perinatal depression needs specialist care.

Doses studied and product quality

Doses studied, not recommendations:

• WFSBP/CANMAT summarised the evidence as 1–10 billion colony-forming units (CFU) a day of varying strains (Sarris 2022).
• London pilot: a multistrain ADM Protexin product, 8 billion CFU a day for 8 weeks alongside an antidepressant (product name and strain count not confirmed in our check) (Nikolova 2023).
• Akkasheh: L. acidophilus, L. casei and B. bifidum, each 2 × 10⁹ CFU/g, for 8 weeks (Akkasheh 2016).
• Basel: Vivomixx, an eight-strain product, 900 billion CFU a day for about one month alongside usual treatment (Schaub 2022).

Because results are strain-specific, a dose from one trial does not transfer to a different product.

Product quality problems

A Slovenian study of supplements sold in pharmacies (Mohar Lorbeg 2021) found that labelled species did not always match the live bacteria in the product, including mislabelling of Lacticaseibacillus casei. We could not confirm the sample size or exact counts from the text we accessed. NCCIH also warns that some products contain organisms not on the label (NCCIH). Even a strain with good trial data cannot help if the product does not deliver it alive.

Follow the money: who makes it and who funded the evidence

Who makes and sells the products tested

• Lallemand Health Solutions (Canada): makes the L. helveticus R0052 + B. longum R0175 combination, marketed as Cerebiome, and describes a research programme on it dating back to 2008 (NutraIngredients 2020). We could not verify Lallemand's ownership structure.
• ADM Protexin (UK, owned by ADM, US): Archer Daniels Midland bought Somerset-based Probiotics International Ltd, maker of Bio-Kult, for £185 million ($243.5 million) in 2018 (NutraIngredients 2018).
• Yakult Honsha (Japan): makes the L. casei Shirota drink; its researchers wrote the exam-stress trials (Takada 2016).
• Nestlé (Switzerland): Nestlé Research Center scientists co-authored the B. longum NCC3001 trial (Pinto-Sanchez 2017).
• Winclove Probiotics (Netherlands): a Winclove employee co-authored both cognitive-reactivity trials (Steenbergen 2015; Chahwan 2019).
• Mendes SA (Switzerland): supplied Vivomixx for the Basel trial (Schaub 2022).
• Chr. Hansen / Novonesis (Denmark): one of the largest probiotic-strain suppliers. It merged with Novozymes on 29 January 2024 to form Novonesis, with about EUR 3.7 billion in pro forma revenue (BioSpace 2024). We did not find a Chr. Hansen-funded mood trial among the studies we reviewed.

Who funded the trials and reviews

The independent syntheses were publicly funded: Liu 2019 by the US National Institute of Mental Health (PubMed grant record), and Asad 2025 by the UK NIHR, a government research funder, with no conflicts declared (Asad 2025). The "add-on" case rests heavily on the King's College London group. Its PhD studentship was an MRC Industrial CASE award with ADM Protexin as partner, and three senior authors received ADM Protexin research grants (Nikolova 2021; Nikolova 2023). Their first review in 2019 declared no conflicts (Nikolova 2019). Lallemand is reported to have funded one trial (Romijn 2017) and supplied products to another (NutraIngredients 2020). The strain-specific meta-analysis states it had no external funding, although the journal record lists Sardinian regional and Italian national (PRIN) grants to one author (Rahmannia 2024). The guideline's lead author declared payments from many nutraceutical companies (Sarris 2022).

Some evidence cuts against a simple "industry bias" story, but we could not confirm every trial's outcome from full text, so that view is provisional. Still, almost every product with positive mood data has a commercial owner involved in its research, and no large, fully independent, adequately powered trial exists.

Documented integrity events

• EU claims: EFSA has not given a positive opinion to any probiotic health-claim application, and the European Ombudsman backed the Commission's restrictive view in December 2024 (FSAI; FoodNavigator 2025).
• US advertising: Dannon's 2010 settlement with the FTC and $21 million settlement with 39 states over exaggerated probiotic claims (FTC 2010). This concerned digestion and immunity claims, not mood.
• US infant products: FDA warning letters in October 2023 to Abbott (Similac Probiotic Tri-Blend) and Infinant Health (Evivo with MCT Oil) (FDA 2023).

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Probiotics: clinical evidence, strains, doses and safety
• Stress, anxiety and depression supplements: evidence and safety
• Major depressive disorder: assessment, treatment evidence and safety
• Treatment-resistant depression
• Generalised anxiety disorder: treatments, supplement evidence and safety
• Irritable bowel syndrome (IBS): diagnosis, treatment and supplement evidence
• Gut health and IBS: evidence, options and safety
• Galactooligosaccharides (GOS): bowel function, IBS and safety
• Saffron for depression
• Omega-3: fish oil, krill and algal forms
• St John's wort vs saffron vs SSRIs for depression

Frequently asked questions

Sources and funding notes

What we could not verify: the exact probiotic effect sizes in Liu 2019 (the overall, anxiety and subgroup figures) and the funding statement for Liu 2019; the Romijn 2017 outcome figures and its funding and supply statements (full text blocked); the Nikolova 2023 effect sizes, product name and company-input details (the lead author's ADM Protexin grants were confirmed); the Schaub 2022 sample size and drop-out figures; the Vinayagamoorthy 2023 pooled counts and mortality; the Mohar Lorbeg 2021 sample size and exact counts; the Kazemi 2019 exact scores and p-value; and the funding statements of Akkasheh 2016, Pinto-Sanchez 2017, Chahwan 2019, Steenbergen 2015, Takada 2016, the Messaoudi 2011 studies and Bafeta 2018. Asad 2025 reports 23 trials, with 1,401 participants in the abstract and 1,280 in the results section. The reported WFSBP/CANMAT 2022 quotations were not re-checked against the full text. We did not verify Chahwan 2019 or Steenbergen 2015 figures in this review. The NIH Office of Dietary Supplements probiotics fact sheet blocked automated access, so we did not use it. We did not independently count EFSA's probiotic claim rejections; we rely on the Irish regulator's statement and the European Commission's reply to the European Ombudsman. The Dannon settlement and FDA 2023 statements were checked against the source pages. The Novonesis revenue figure is a pre-merger pro forma estimate.

1. NCCIH. Probiotics: What You Need To Know (updated August 2019) — US regulator-level fact sheet; funded by US Congress via NIH.
2. FDA. FDA raises concerns about probiotic products sold for use in hospitalized preterm infants (26 October 2023) — US government.
3. FDA. Risk of invasive disease in preterm infants given probiotics — US government safety communication.
4. Food Safety Authority of Ireland. Probiotic health claims — Irish government regulator.
5. FoodNavigator. Why the EU still restricts "probiotics" claims (January 2025) — trade press; advertising-funded.
6. FTC. Dannon agrees to drop exaggerated health claims (December 2010) — US government.
7. Liu RT et al. Prebiotics and probiotics for depression and anxiety: meta-analysis. Neurosci Biobehav Rev 2019 — NIMH grants.
8. Asad A et al. Prebiotics and probiotics in clinically diagnosed samples. Nutr Rev 2025 — NIHR Oxford Health BRC; no conflicts.
9. Nikolova VL et al. Updated review and meta-analysis of probiotics for clinical depression. J Clin Med 2021 — MRC studentship (ADM Protexin partner); co-author ADM Protexin grants.
10. Nikolova V et al. Gut feeling: RCTs of probiotics for clinical depression. Ther Adv Psychopharmacol 2019 — MRC studentship; no conflicts declared.
11. Nikolova VL et al. Probiotics as adjunctive treatment in depression: pilot RCT. JAMA Psychiatry 2023 — MRC Industrial CASE with ADM Protexin; Freya Green; product from ADM Protexin.
12. Nikolova VL et al. Mechanisms of action of probiotics in depression. J Affect Disord 2025 — MRC, NIHR, King's College London; one author affiliated with ADM Protexin.
13. Nikolova VL et al. Gut microbiota perturbations in psychiatric disorders. JAMA Psychiatry 2021 — MRC, BBSRC and UK Department of Health grants (PubMed grant record).
14. Rahmannia M et al. Strain-specific effects of probiotics on depression and anxiety. Gut Pathog 2024 — funding statement says no external funding; journal record lists Sardinia region and Italian PRIN grants to one author; one author acknowledged a Shahid Beheshti University grant.
15. Sarris J et al. WFSBP/CANMAT clinician guidelines on nutraceuticals and phytoceuticals. World J Biol Psychiatry 2022 — professional societies; lead author declared payments from many nutraceutical companies.
16. Bahji A et al. CANMAT task force recommendations for probiotics, prebiotics, synbiotics and faecal microbiota transplants in adults with MDD. Can J Psychiatry, November 2025 — no external funding; no conflicts declared.
17. Schaub AC et al. Probiotic add-on therapy in depressed patients. Transl Psychiatry 2022 — Basel university and clinic foundations; product supplied by Mendes SA; no competing interests.
18. Romijn AR et al. L. helveticus and B. longum for symptoms of depression. Aust N Z J Psychiatry 2017 — funded by Lallemand Health Solutions.
19. Kazemi A et al. Probiotic and prebiotic vs placebo in MDD. Clin Nutr 2019 — Tehran University of Medical Sciences grant; products from Lallemand (trade press).
20. Akkasheh G et al. Probiotic administration in MDD. Nutrition 2016 — Iranian universities; funding not accessed.
21. Kelly JR et al. L. rhamnosus (JB-1) fails to modulate stress in healthy males. Brain Behav Immun 2017 — Health Research Board, Science Foundation Ireland; one co-author affiliated with Alimentary Health Ltd.
22. Bravo JA et al. Lactobacillus strain regulates emotional behaviour in a mouse via the vagus nerve. PNAS 2011 — University College Cork; animal study.
23. Pinto-Sanchez MI et al. B. longum NCC3001 reduces depression scores in IBS. Gastroenterology 2017 — McMaster University with Nestlé Research Center co-authors.
24. Takada M et al. L. casei strain Shirota relieves stress-associated symptoms. Neurogastroenterol Motil 2016 — Yakult-affiliated authors.
25. Messaoudi M et al. Psychotropic-like properties of a probiotic formulation in rats and humans. Br J Nutr 2011 — ETAP contract laboratory; no funding statement in PubMed record.
26. Steenbergen L et al. Multispecies probiotics and cognitive reactivity to sad mood. Brain Behav Immun 2015 — Leiden/Amsterdam universities with a Winclove co-author.
27. Chahwan B et al. Gut feelings: probiotics for depressive symptoms. J Affect Disord 2019 — University of Technology Sydney with a Winclove co-author.
28. Yelin I et al. Bacterial transmission from probiotic capsule to blood in ICU patients. Nat Med 2019 — NIH, European Research Council, Boston Children's Hospital; no competing interests.
29. Vinayagamoorthy K et al. Epidemiology of Saccharomyces fungemia: systematic review. Med Mycol 2023 — academic; funding not accessed.
30. Bafeta A et al. Harms reporting in RCTs of microbiota-modifying interventions. Ann Intern Med 2018 — no specific funding.
31. Mohar Lorbeg P et al. Evaluation of dietary supplements containing viable bacteria. Front Microbiol 2021 — Slovenian Research Agency.
32. NutraIngredients. Lallemand's psychobiotic combination may boost BDNF levels (October 2020) — trade press.
33. NutraIngredients. ADM expands probiotics reach with $243 million acquisition (June 2018) — trade press.
34. BioSpace. Novozymes and Chr. Hansen combination completed, creating Novonesis (January 2024) — company announcement.
35. NCCIH funding appropriations history — US government.

Educational research reviewed 7 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.