MENU
Healthspan, by Evidence.
  1. Home
  2. サプリメント
  3. Do Probiotic Drinks and Supplements Really Work? The Cases Where They Help Are Surprisingly Narrow [2026 Evidence Review]

Do Probiotic Drinks and Supplements Really Work? The Cases Where They Help Are Surprisingly Narrow [2026 Evidence Review]

サプリメント
bowl of yogurt with a spoon

Do Probiotic Drinks and Supplements Actually Work? The Situations Where They Help Are Surprisingly Narrow [2026 Evidence Edition]

[Level 2 (Moderate)] [Supplements] [Series: Health for Serious Middle-Agers]

TOC

⚠️ Before You Read This Article

This article provides educational information on probiotics (live-microorganism products such as lactobacilli and bifidobacteria), based on peer-reviewed literature (Cochrane reviews, RCTs, and meta-analyses). It is not a substitute for medical advice. If you are immunosuppressed, critically ill, or have a central venous catheter in place, be sure to consult your physician before using probiotics.

Key Points of This Article (3-Line Summary)

  • The only use supported with “moderate certainty” is prevention of antibiotic-associated diarrhea (AAD) and C. difficile infection (specific strains, 5 billion CFU or more). However, the largest RCT in hospitalized patients aged 65 and over found no effect.
  • For healthy people, “cold prevention,” “immune boosting,” “mood improvement,” and “weight loss” show either small effects (colds −24% with low certainty, depression d=−0.24, body weight −0.6 kg) or rest mainly on small, industry-funded studies.
  • “Lactic acid bacteria” cannot be lumped together (strain specificity). In healthy people the gut microbiota barely changes, and a Cell 2018 study provides counter-evidence that taking probiotics after antibiotics actually delays recovery.

1. Where the Evidence Supports Use: Antibiotic-Associated Diarrhea

The most solid evidence for probiotics is in preventing diarrhea while taking antibiotics.

  • Pediatric AAD (Guo, Goldenberg et al., Cochrane 2019; 33 RCTs, 6,352 participants): reduced incidence, NNT approximately 9 (moderate certainty). Effective strains were Lactobacillus rhamnosus GG and Saccharomyces boulardii, at 5–40 billion CFU per day.
  • C. difficile-associated diarrhea (Goldenberg et al., Cochrane 2017): incidence with concurrent antibiotics 1.5% vs 4.0%, RR 0.40, NNT 40 (moderate). However, no significant difference in populations with a baseline risk below 5%.

Here there is an important piece of counter-evidence. The PLACIDE trial (Allen et al., Lancet 2013; hospitalized patients aged 65 and over, N=2,981), the largest RCT in the UK, found that a combined lactobacillus-plus-bifidobacterium preparation reduced neither AAD nor C. difficile (approximately 10% vs 10%). For readers in their 50s and 60s, the fact that “it did not work in older adults” carries real weight; understand that “it works” is based mainly on data from children and younger adults.

In pediatric acute gastroenteritis as well, Freedman et al. (NEJM 2018, N=886) found no difference in diarrhea severity or duration with LGG.

2. Irritable Bowel Syndrome (IBS) and Constipation

Ford et al. (Aliment Pharmacol Ther 2018; 53 RCTs, 5,545 participants) showed a slight benefit for overall symptoms, but concluded that “only specific strains or combinations” were involved and that “definitive conclusions cannot be drawn.” In the 2023 update (82 trials, 10,332 participants), GRADE certainty was low to very low. The American Gastroenterological Association (AGA 2020) does not recommend use outside of clinical trials for IBS or general digestive symptoms.

In practical terms, a reasonable approach is to “try for 4–8 weeks and stop if it does not help.”

3. Cold Prevention and “Immunity”

Cochrane (Zhao et al. 2022, CD006895): reduced the proportion of people experiencing at least one acute upper respiratory tract infection by about 24% (low certainty), reduced the proportion experiencing three or more episodes (moderate), and reduced antibiotic use. However, many trials were industry-funded and small, and the authors themselves noted that “conclusions are limited.”

As a Japanese example, Shida et al. (Eur J Nutr 2017; male office workers, N=96) reported upper respiratory tract infections of 22% vs 53% after 12 weeks of a L. casei strain Shirota drink, but this was a study by the manufacturer’s research institute, small, and men only, with a clear conflict of interest. Independent replication is scarce.

4. Mental Health (Psychobiotics), Body Weight, and Cholesterol

  • Depression and anxiety: Liu et al. (Neurosci Biobehav Rev 2019, 34 trials) found depression d=−0.24 and anxiety d=−0.10 (both “small”). The effect almost disappears in healthy people, while in clinical populations d=−0.45. There is little basis for healthy people in their 50s and 60s to take probiotics for “mood improvement.”
  • Body weight: Borgeraas et al. (Obes Rev 2018, 15 RCTs) found −0.60 kg. Statistically significant, but of almost no clinical relevance.
  • Lipids: Shimizu et al. (PLoS One 2015, 11 RCTs) found LDL −0.22 mmol/L (approximately −8.5 mg/dL). Small compared with statins (−30 to 50%), but potentially a useful adjunct for mildly elevated cholesterol.

5. “Lactic Acid Bacteria” Cannot Be Lumped Together: Strain Specificity and Individual Differences in Colonization

Effects should be evaluated at the strain level and cannot be extrapolated to “lactobacilli in general” or “bifidobacteria in general.” The industry’s own criticism of PLACIDE, that “the strains used had no evidence for AAD,” itself illustrates this principle.

Furthermore, Zmora et al. (Cell 2018) administered an 11-strain preparation to healthy volunteers and sampled the gut mucosa directly by endoscopy. Although the bacteria passed through alive, mucosal colonization varied by person, site, and strain, and participants split into “colonizers” and “resisters.” Finding the bacteria in stool does not mean they have colonized the mucosa. Likewise, a review of 7 RCTs in healthy adults by Kristensen et al. (Genome Med 2016) found no significant change in the composition of the fecal microbiota.

6. “Restoring Your Gut After Antibiotics” May Backfire

Suez et al. (Cell 2018) divided 21 healthy volunteers who had received antibiotics into three groups: (1) spontaneous recovery, (2) an 11-strain probiotic, and (3) autologous fecal transplant. The result: the probiotic group showed markedly delayed recovery of the gut microbiota, which was still incomplete about 5 months later. The autologous fecal transplant group recovered within days. Secreted factors from the lactobacilli were suppressing the native bacteria.

This was a single product, a small study, and not a clinical-symptom outcome, but it is a serious piece of counter-evidence to the popular advice to “reset your gut with probiotics after antibiotics.” Think of taking probiotics “at the same time” as antibiotics (AAD prevention) and taking them “after finishing” as two different things.

7. Safety: Fine for Healthy People, but the Exceptions Can Be Life-Threatening

Serious adverse events are rare in healthy people. However, in the PROPATRIA trial (Besselink et al., Lancet 2008; predicted severe acute pancreatitis, N=296), the multi-species probiotic group had mortality of 16% vs 6%, with 9 cases of bowel ischemia. Yelin et al. (Nat Med 2019) genomically confirmed 22 cases in which L. rhamnosus GG originating from capsules caused bloodstream infections in ICU patients. Case reports of bacteremia continue to accumulate in immunosuppressed patients, those with central venous catheters, the critically ill, and those with short bowel syndrome.

8. How to Read Japan’s FOSHU (Foods for Specified Health Uses) Label

FOSHU (Foods for Specified Health Uses, Japan’s regulated health-claim label) is based on individual review by the Consumer Affairs Agency, but most of the supporting trials are company-run, small (N=a few dozen), and short-term, reaching only low to moderate certainty under international GRADE criteria. The claim “helps maintain a healthy gut” primarily refers to improvements in bowel movement frequency and stool consistency; disease-prevention effects are outside the permitted scope. Note that appeals to “immunity,” “cognition,” and the like are often advertising language rather than FOSHU-approved claims.

9. Conditional Conclusions

Situations where benefit is likely – Prevention of diarrhea and C. difficile while taking antibiotics (LGG or S. boulardii, 5 billion CFU/day or more, started within 2 days of beginning antibiotics). However, negative in hospitalized patients aged 65 and over. – Adjunct for mildly elevated cholesterol (LDL reduction of about 8 mg/dL). – Relief of IBS symptoms with specific strains (low certainty; give up after 4–8 weeks if no benefit).

Situations where it does not work or evidence is insufficient – Cold prevention or “immune boosting” in healthy people. – Mood improvement or weight loss in healthy people. – “Gut reset” after finishing antibiotics.

Situations to avoid – Critical illness, ICU, immunosuppression, central venous catheter, short bowel syndrome.

It is more accurate to think of fermented foods such as yogurt and natto (see our existing articles) separately from “supplements and drinks for taking in bacteria.” The former have consistent observational data as part of a dietary pattern; the latter have shown effects only when the strain and the situation are narrowly defined.

References and Sources

  • Guo Q, Goldenberg JZ, et al. (2019). Probiotics for the prevention of pediatric antibiotic-associated diarrhea. Cochrane Database Syst Rev CD004827.
  • Goldenberg JZ, et al. (2017). Probiotics for the prevention of Clostridium difficile-associated diarrhea. Cochrane Database Syst Rev CD006095.
  • Allen SJ, et al. (2013). Lactobacilli and bifidobacteria in the prevention of antibiotic-associated diarrhoea and Clostridium difficile diarrhoea in older inpatients (PLACIDE). Lancet 382:1249-1257.
  • Freedman SB, et al. (2018). Multicenter trial of a combination probiotic for children with gastroenteritis. N Engl J Med 379:2015-2026.
  • Ford AC, et al. (2018). Systematic review with meta-analysis: the efficacy of prebiotics, probiotics, synbiotics and antibiotics in IBS. Aliment Pharmacol Ther 48:1044-1060.
  • Zhao Y, Dong BR, Hao Q. (2022). Probiotics for preventing acute upper respiratory tract infections. Cochrane Database Syst Rev CD006895. PMID: 36001877
  • Shida K, et al. (2017). Daily intake of fermented milk with Lactobacillus casei strain Shirota reduces the incidence and duration of upper respiratory tract infections. Eur J Nutr 56:45-53. PMID: 26419583
  • Liu RT, et al. (2019). Prebiotics and probiotics for depression and anxiety: meta-analysis. Neurosci Biobehav Rev 102:13-23.
  • Borgeraas H, et al. (2018). Effects of probiotics on body weight. Obes Rev 19:219-232. PMID: 29047207
  • Shimizu M, et al. (2015). Meta-analysis: effects of probiotic supplementation on lipid profiles. PLoS One 10:e0139795.
  • Zmora N, et al. (2018). Personalized gut mucosal colonization resistance to empiric probiotics. Cell 174:1388-1405. PMID: 30193112
  • Suez J, et al. (2018). Post-antibiotic gut mucosal microbiome reconstitution is impaired by probiotics. Cell 174:1406-1423. PMID: 30193113
  • Besselink MG, et al. (2008). Probiotic prophylaxis in predicted severe acute pancreatitis (PROPATRIA). Lancet 371:651-659.
  • Yelin I, et al. (2019). Genomic and epidemiological evidence of bacterial transmission from probiotic capsule to blood in ICU patients. Nat Med 25:1728-1732.

Evidence current as of October 2026. We recommend verifying details such as the RR/NNT in Ford 2018 and the bibliographic details of AGA 2020 against the original sources.

サプリメント
Let's share this post !

Author of this article

masa_evidageのアバター masa_evidage

関連記事

No related articles were found yet.

TOC