- Design
- systematic review and meta-analysis of 13 randomised trials and one interventional study; subgroup meta-analysis for pooling
- Population
- adults aged 60 and over, healthy or cognitively impaired, across 14 studies of probiotics, prebiotics or synbiotics
- Primary outcome
- change in cognitive function on standardised instruments
- Effect
- pooled global cognition mean difference 1.28 points (95% CI 0.15-2.40)
Fourteen studies — 13 randomised trials and one interventional study — of prebiotic, probiotic or synbiotic supplementation in adults aged 60 and over were reviewed, with cognitive outcomes measured by standard instruments. Faecal transplant studies and head-to-head comparisons without placebo were excluded.
Nine studies reported a significant improvement in at least one cognitive domain, most often global cognition, memory, attention or executive function. A subgroup meta-analysis produced a pooled improvement in global cognition of 1.28 points (95% CI 0.15-2.40).
That number needs placing. On a 30-point instrument such as the Mini-Mental State Examination, a difference of around 1 to 3 points is usually regarded as the minimum a clinician could detect, and the confidence interval here runs from 0.15 — which is nothing — to 2.40. The lower half of that interval describes a change no patient or family would notice. The authors' conclusion that this holds 'huge promise' is considerably stronger than the interval supports, and only a subgroup of the studies could be pooled at all.
The argument that these products are already available and low-risk is fair as far as it goes, and it is why this is worth knowing rather than dismissing: a patient taking a probiotic is not doing themselves harm. But 'available and harmless' is a reason not to object, not a reason to recommend, and recommending a supplement on this evidence spends credibility that is needed for the interventions with better support — exercise, hearing correction, blood pressure control and social contact.
- Do not recommend probiotics or prebiotics for cognition; the pooled gain includes changes too small to detect clinically
- Do not object where a patient is already taking one — no harm signal emerged and the products are low-risk
- Direct the conversation to interventions with stronger support: exercise, hearing correction, blood pressure control and social contact
- Note that only a subgroup of studies could be pooled, so the meta-analytic estimate rests on part of the evidence
- Ask what else the patient is spending on supplements; cost matters more than the supplement does
Why it matters
Supplements marketed for brain health now cite meta-analyses, and this is what one of those meta-analyses actually contains.
Don't overread it
A statistically significant pooled difference whose interval nearly reaches zero does not establish a clinically meaningful benefit.
The statistics, in plain English
A mean difference of 1.28 points with an interval from 0.15 to 2.40 spans the range between clinically invisible and marginally detectable, so the honest summary is that the size of any benefit is unknown. The estimate comes from a subgroup of the 14 included studies rather than all of them, which means it was pooled where pooling was possible rather than where it was planned — a route by which optimistic estimates commonly arise. Fourteen studies of three different intervention types in populations ranging from healthy to cognitively impaired is not one question being answered.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for geriatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free