Effect of Probiotic Supplementation on Cognitive Function and Metabolic Status in Alzheimer's Disease: A Randomized, Double-Blind and Controlled Trial

The researchers tested whether daily consumption of probiotic-enriched milk could improve cognitive function and metabolic health in people with Alzheimer's disease (AD). The intervention was 200 mL/day of milk containing four specific bacterial strains: Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium bifidum, and Lactobacillus fermentum — each at a dose of 2 × 10⁹ colony-forming units (CFU) per gram. The comparator was plain milk (200 mL/day) with no probiotics. The primary outcome was cognitive function measured by the Mini-Mental State Examination (MMSE). Secondary outcomes included markers of oxidative stress (malondialdehyde, total antioxidant capacity, glutathione, nitric oxide), inflammation (high-sensitivity C-reactive protein), and metabolic status (fasting blood glucose, insulin, insulin resistance, beta-cell function, and blood lipids including triglycerides, total cholesterol, LDL, and HDL).

The study included 60 patients diagnosed with Alzheimer's disease, aged 60–95 years, residing at two welfare organizations in Kashan and Esfahan, Iran. Diagnosis followed standard NINDS-ADRDA criteria and revised National Institute on Aging-Alzheimer's Association criteria. The sample was predominantly female: 48 women and 12 men (24 women and 6 men in each group). Patients with metabolic disorders, chronic infections, or other clinically relevant conditions besides AD were excluded. Anyone who had consumed probiotic supplements within 6 weeks prior to the study, or who regularly ate probiotic yogurt, kefir, or other fermented foods, was also excluded. Four patients died during the study (2 per group), leaving 52 completers, but all 60 were included in the final intention-to-treat analysis.

Cognitive function was assessed using the Mini-Mental State Examination (MMSE), a 30-point questionnaire widely used to screen for cognitive impairment. Scores range from 0 (severe impairment) to 30 (normal cognition). A score of 24 or higher is generally considered normal; 18–23 indicates mild impairment; and below 18 indicates moderate-to-severe impairment. The MMSE tests orientation, attention, memory, language, and visuospatial skills.

Blood biomarkers were measured from 12-hour fasting blood samples collected at baseline and after 12 weeks:

• Malondialdehyde (MDA) — a marker of oxidative stress, measured via the thiobarbituric acid reactive substance method
• High-sensitivity C-reactive protein (hs-CRP) — a marker of systemic inflammation, measured by ELISA
• Total antioxidant capacity (TAC) — measured using the ferric reducing antioxidant power method
• Total glutathione (GSH) — an antioxidant, measured by the Beutler method
• Nitric oxide (NO) — measured by the Griess method
• Fasting plasma glucose (FPG) and serum lipids (triglycerides, total cholesterol, LDL, HDL) — measured using commercial kits
• Serum insulin — measured by ELISA
• HOMA-IR (homeostatic model assessment for insulin resistance) — calculated from fasting glucose and insulin
• HOMA-B (beta-cell function) — calculated from fasting glucose and insulin
• QUICKI (quantitative insulin sensitivity check index) — calculated from fasting glucose and insulin

Dietary intake was assessed using 3-day food records collected at baseline, weeks 3, 6, 9, and 12, analyzed with Nutritionist IV software.

Study design: This was a randomized, double-blind, placebo-controlled clinical trial (RCT) — the gold standard for testing causal effects of an intervention.

Randomization: Participants were matched for disease severity based on gender, BMI, and age at baseline, then randomly assigned to either the probiotic or control group using computer-generated random numbers. The allocation sequence was concealed from researchers and participants until final analysis was complete.

Blinding: The study was double-blind, meaning neither the participants nor the researchers assessing outcomes knew which group received the probiotic milk versus plain milk. This is critical because expectation effects can influence both subjective cognitive performance and biological markers.

Duration: The intervention lasted 12 weeks, which is a reasonable duration for observing changes in gut microbiota composition and downstream metabolic effects. However, 12 weeks is relatively short for assessing cognitive changes in a progressive neurodegenerative condition like Alzheimer's.

Statistical approach: The primary analysis used intention-to-treat (ITT), meaning all 60 randomized participants were included in the final analysis regardless of whether they completed the study. Missing data were handled using last-observation-carried-forward (LOCF), which assumes the participant's condition remained stable after dropout — a conservative but potentially biased approach if dropouts were deteriorating. Independent samples t-tests were used to compare changes between groups, and ANCOVA was used to adjust for baseline values, age, and BMI. Log transformation was applied to non-normally distributed variables (FPG, insulin, HOMA-IR, hs-CRP). The sample size calculation (25 per group, inflated to 30 to account for 5 dropouts) was based on detecting a 1.1-point difference in MMSE with 80% power at α = 0.05.

What this design can prove: Because of randomization, blinding, and a control group, this study can establish that the probiotic intervention caused the observed changes in MMSE scores and metabolic markers — not that these changes were due to placebo effects, natural fluctuations, or confounding variables.

What this design cannot prove:

• It cannot prove the mechanism — whether cognitive improvements were directly due to gut-brain signaling, reduced inflammation, improved insulin sensitivity, or some combination
• It cannot prove long-term effects beyond 12 weeks
• It cannot prove generalizability to younger patients, those living at home (not in institutions), or those with different disease severities
• The LOCF method for missing data may underestimate or overestimate effects if dropouts were systematically different from completers

Methodological weaknesses:

• Small sample size (n=60, with 4 deaths)
• No assessment of gut microbiota composition — we don't know if the probiotics actually colonized or changed the microbiome
• No measure of compliance beyond self-report and food records
• The MMSE is a relatively crude cognitive screening tool; more sensitive neuropsychological tests might have detected different patterns
• The control group received plain milk, not a placebo matched for taste, texture, or caloric content — though blinding was claimed
• Industry funding or conflicts of interest are not explicitly stated, but the probiotics were produced by a commercial company (Tak Gen Zist Pharmaceutical Company)

Primary outcome — Cognitive function (MMSE):

• Probiotic group: MMSE score increased by 27.90% ± 8.07 (from baseline to 12 weeks)
• Control group: MMSE score decreased by 5.03% ± 3.00
• Between-group difference: p < 0.001 (highly significant)
• This means the probiotic group improved by roughly 28% while the control group declined by about 5%

Secondary outcomes — Inflammation and oxidative stress:

• Malondialdehyde (MDA) — decreased by 22.01% ± 4.84 in the probiotic group vs. increased by 2.67% ± 3.86 in controls (p < 0.001)
• High-sensitivity C-reactive protein (hs-CRP) — decreased by 17.61% ± 3.70 in the probiotic group vs. increased by 45.26% ± 3.50 in controls (p < 0.001)
• Total antioxidant capacity (TAC) — no significant difference between groups
• Total glutathione (GSH) — no significant difference between groups
• Nitric oxide (NO) — no significant difference between groups

Secondary outcomes — Metabolic markers:

• HOMA-IR (insulin resistance) — increased by 28.84% ± 13.34 in the probiotic group vs. increased by 76.95% ± 24.60 in controls (p = 0.002) — meaning both groups got worse, but the probiotic group worsened significantly less
• HOMA-B (beta-cell function) — increased by 3.45% ± 10.91 in the probiotic group vs. increased by 75.62% ± 23.18 in controls (p = 0.001) — again, both increased, but the control group increased much more (which may reflect compensatory insulin secretion in response to worsening insulin resistance)
• QUICKI (insulin sensitivity) — decreased by 1.83 ± 1.26 in the probiotic group vs. decreased by 4.66 ± 1.70 in controls (p = 0.006) — both groups lost insulin sensitivity, but the probiotic group lost less
• Triglycerides — decreased by 20.29% ± 4.49 in the probiotic group vs. decreased by 0.16% ± 5.24 in controls (p = 0.003)
• Fasting plasma glucose — no significant difference between groups
• Total cholesterol, LDL, HDL — no significant differences between groups

No significant effects were found on: total antioxidant capacity, glutathione, nitric oxide, fasting glucose, total cholesterol, LDL, or HDL.

The most striking finding is the cognitive improvement: the probiotic group's MMSE score improved by roughly 28% over 12 weeks, while the control group declined by about 5%. To put this in context, a typical Alzheimer's patient might decline by 2–4 points on the MMSE per year. An improvement of 28% over 12 weeks is unusually large — it would be equivalent to a patient moving from moderate impairment (MMSE ~15) to mild impairment (MMSE ~19) in just 3 months. However, the absolute MMSE scores at baseline and post-treatment are not reported in the abstract (only percent changes), so the clinical significance depends on where patients started.

For inflammation: hs-CRP dropped by nearly 18% in the probiotic group while rising 45% in controls — a net difference of about 63 percentage points. This is a large effect. For context, hs-CRP levels above 3 mg/L indicate high cardiovascular risk; a 18% reduction could move someone from high to moderate risk.

For oxidative stress: MDA (a marker of lipid peroxidation) dropped by 22% in the probiotic group while rising slightly in controls. This is a moderate-to-large effect.

For insulin resistance: both groups got worse, but the probiotic group's HOMA-IR increased by only 29% compared to 77% in controls — a net benefit of about 48 percentage points. This suggests probiotics may slow the metabolic deterioration associated with AD.

For triglycerides: a 20% reduction in the probiotic group vs. essentially no change in controls is clinically meaningful, as elevated triglycerides are a cardiovascular risk factor.

What the authors acknowledge:

• The study was relatively short (12 weeks)
• The sample size was small (n=60)
• They did not measure gut microbiota composition directly
• The MMSE is a relatively simple cognitive screening tool
• They did not assess whether the probiotics actually colonized the gut

What a critical reader would note:

• No microbiome analysis: Without measuring gut bacteria, we cannot confirm that the probiotics altered the microbiome — the effects could be due to other components of the milk or unknown factors
• High dropout rate from deaths: 4 patients died (6.7%), which is concerning in a 12-week study. The LOCF method for handling missing data assumes no change after dropout, but if these patients were deteriorating, this could bias results
• Population specificity: All participants were institutionalized in welfare organizations in Iran, aged 60–95, predominantly female. Results may not generalize to younger, community-dwelling, or male patients
• No placebo control for taste/texture: The control group received plain milk, not a probiotic-free milk matched for taste. If the probiotic milk tasted different, blinding could be compromised
• No assessment of diet changes: While food records were collected, the authors don't report whether the probiotic group changed their diet in other ways (e.g., eating more fermented foods)
• Multiple comparisons: Many outcomes were tested (MMSE, MDA, hs-CRP, TAC, GSH, NO, FPG, insulin, HOMA-IR, HOMA-B, QUICKI, triglycerides, total cholesterol, LDL, HDL). Without correction for multiple comparisons, some significant findings could be due to chance
• Industry involvement: The probiotics were produced by a commercial company (Tak Gen Zist Pharmaceutical Company), though no explicit conflict of interest is stated
• No long-term follow-up: We don't know if cognitive improvements persisted after stopping probiotics
• No assessment of functional outcomes: The MMSE measures cognitive performance, but we don't know if patients had better quality of life, daily functioning, or caregiver burden

For someone running their own n=1 experiment to test whether probiotics might support cognitive function:

What to test

• Specific intervention: A multi-strain probiotic containing Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium bifidum, and Lactobacillus fermentum — each at a dose of at least 2 × 10⁹ CFU per serving. Look for a commercial product that matches these strains and doses. Alternatively, you could test a high-quality multi-strain probiotic with similar species (e.g., Lactobacillus and Bifidobacterium combinations).
• Dose: 200 mL of probiotic milk daily (or equivalent capsule form). If using capsules, aim for a total of ~8 × 10⁹ CFU per day (2 × 10⁹ of each of four strains).
• Comparator: Your baseline (no probiotic) for at least 4 weeks before starting, or a placebo (e.g., plain milk or inert capsules) if you can source them.

Minimum meaningful duration

• At least 12 weeks — this study showed effects at 12 weeks. Gut microbiota changes can occur within 1–2 weeks, but cognitive and metabolic effects may take longer to manifest.
• Consider running for 16–24 weeks to see if effects plateau, continue, or reverse.

What to measure

• Cognitive function: Use a validated online cognitive assessment tool that tests memory, attention, and executive function. The MMSE is not available for self-administration, but alternatives include:
  • Cogstate (online cognitive testing battery)
  • Lumosity or BrainHQ (though these are training tools, not diagnostic)
  • Self-administered MoCA (Montreal Cognitive Assessment) — available online, but best done with a partner
  • Simple reaction time and working memory tests (many free apps available)
• Inflammation: If you have access to blood testing, measure hs-CRP (high-sensitivity C-reactive protein) at baseline and end of study. This costs ~$30–50 per test.
• Oxidative stress: MDA (malondialdehyde) is not commonly available in standard labs, but you could measure urinary 8-OHdG (a DNA oxidation marker) via mail-in kits.
• Metabolic markers: Fasting glucose and fasting insulin (to calculate HOMA-IR). Triglycerides are also easy to measure. These are standard blood tests (~$20–50 each).
• Subjective measures: Keep a daily log of:
  • Mental clarity (1–10 scale)
  • Memory lapses (count per day)
  • Energy levels (1–10 scale)
  • Digestive symptoms (bloating, gas, stool consistency)
  • Sleep quality (hours and subjective rating)

Key confounds to control for