Clinical, gut microbial and neural effects of a probiotic add-on therapy in depressed patients: a randomized controlled trial

The researchers tested whether a high-dose, multi-strain probiotic supplement (Vivomixx®) taken daily for 31 days, in addition to each patient's usual depression treatment (treatment-as-usual, TAU), would reduce depressive symptoms compared to a placebo (maltose powder) also added to TAU. They measured depressive symptoms using a clinician-rated scale (Hamilton Depression Rating Scale, HAM-D) as the primary outcome. Secondary outcomes included self-reported depression (Beck Depression Inventory, BDI), anxiety (State-Trait Anxiety Inventory, STAI), gastrointestinal symptoms (Gastrointestinal Symptom Rating Scale, GSRS), changes in gut bacteria composition (via 16S rRNA gene sequencing of stool samples), and brain activity (via functional MRI while viewing emotional faces). Assessments occurred at three time points: baseline (before the intervention), immediately after the 31-day intervention, and at a follow-up four weeks after the intervention ended.

**Sample size:** 60 patients were enrolled; 47 completed the intervention (21 in the probiotic group, 26 in the placebo group). The intention-to-treat (ITT) analysis included all 60; the modified ITT (mITT) included the 47 completers.

**Population:** Adult inpatients (hospitalized) at the University Psychiatric Clinics in Basel, Switzerland, with a current depressive episode (diagnosed as F31.3–F34 under ICD-10 criteria, which includes unipolar depression and bipolar depression, current episode depressed).

**Inclusion criteria:** HAM-D score >7 (mild depression or higher), age ≥18 years, receiving treatment-as-usual for depression.

**Exclusion criteria:** Immunosuppression, dietary restrictions, acute infectious diseases, pregnancy, breastfeeding, comorbid psychiatric disorders (addiction, bipolar disorder, schizophrenia).

**Demographics:** Mean age was approximately 40 years (range not given); about 60% were female. Groups were balanced for age, sex, BMI, and baseline depression severity.

**Setting:** Inpatient psychiatric hospital; patients who were discharged during the 31-day intervention were instructed to continue taking the supplement at home.

**Primary outcome:** Hamilton Depression Rating Scale (HAM-D, 17-item version). A clinician-administered interview; scores range 0–52, with higher scores indicating more severe depression. A score of 0–7 is considered normal, 8–13 mild, 14–18 moderate, 19–22 severe, and ≥23 very severe.

**Secondary clinical outcomes:** Beck Depression Inventory (BDI, self-report, 0–63 scale, higher = worse), State-Trait Anxiety Inventory – State version (STAI-1, self-report, 20–80 scale, higher = more anxiety), Gastrointestinal Symptom Rating Scale (GSRS, self-report, 0–45 scale, higher = more GI distress).

**Gut microbiota:** Stool samples were collected at baseline and post-intervention. DNA was extracted, and the V4 region of the 16S ribosomal RNA gene was sequenced. Quantitative microbiome profiling (including flow cytometry to measure total microbial load) was used to estimate absolute abundances of bacterial taxa. Enterotyping classified samples into community types (Prevotella, Bacteroides 2, or Ruminococcus).

**Brain imaging:** Functional MRI (fMRI) was performed at baseline and post-intervention. Participants viewed blocks of fearful, happy, and neutral faces (from the Ekman faces set) and a control condition (scrambled faces). Brain activation was measured in regions of interest: amygdala, putamen, caudate, and nucleus accumbens.

**Timing:** Baseline assessment (week 0), post-intervention (week 4, after 31 days of supplementation), and follow-up (week 8, four weeks after stopping the intervention). Brain imaging was only done at baseline and post-intervention, not at follow-up.

**Design:** Double-blind, placebo-controlled, parallel-group randomized controlled trial (RCT).

**Randomisation:** Patients were randomly allocated to probiotic or placebo groups. The method of randomisation (e.g., computer-generated sequence, sealed envelopes) is not detailed in the abstract, but the full paper likely used a standard random number generator.

**Blinding:** Double-blind — neither the patients, the clinicians administering the HAM-D, the nursing staff dispensing the supplements, nor the researchers analyzing the data knew which group a patient was in. The placebo (maltose) was matched for color, shape, size, packaging, smell, and taste.

**Duration:** The intervention lasted 31 days. Follow-up was 4 weeks after the intervention ended (total study duration 8 weeks per participant). This is a relatively short intervention for a psychiatric trial; most antidepressant trials last 8–12 weeks.

**Statistical approach:** Mixed-effects models (with random intercepts) were used to analyze HAM-D scores over time (baseline, post-intervention, follow-up) by group (probiotic vs. placebo). An ANOVA (type III) was run on the model. Post-hoc tests compared change scores between groups. Effect sizes (Cohen's d) were calculated. Analyses were run on both the intention-to-treat (ITT) sample (all 60 randomized) and the modified ITT (mITT) sample (47 completers). Potential confounders (sex, age, BMI, medication) were included as covariates in sensitivity analyses.

**What this design can prove:** A double-blind RCT is the gold standard for establishing causality. If the probiotic group improves more than the placebo group, and blinding was successful, you can conclude that the probiotic *caused* the improvement (relative to placebo). The inclusion of a follow-up period allows testing whether effects persist after stopping the intervention.

**What this design cannot prove:**

It cannot prove that the probiotic works *without* concurrent antidepressant medication, because all patients were on treatment-as-usual (which included antidepressants, psychotherapy, etc.). This is an "add-on" trial.

It cannot prove that the probiotic is better than an active comparator (e.g., another probiotic, a dietary change, or a different dose).

The short duration (31 days) cannot rule out that effects would be larger or smaller with longer treatment.

The small sample size (47 completers) means the study is underpowered to detect small effects or to reliably examine subgroups (e.g., by sex, age, or medication type).

The single-center design limits generalizability to other hospitals, countries, or outpatient settings.

**Major methodological weaknesses:**

The inclusion criteria were changed mid-study: initially only patients with severe depression (HAM-D >24) were eligible, but this was relaxed to HAM-D >7 (mild depression) to improve recruitment. This change was not pre-registered and could introduce bias.

The study was funded by the manufacturer of the probiotic (Mendes SA, Lugano, Switzerland, which produces Vivomixx®). Industry funding is a known source of bias in nutrition research.

The placebo was maltose, a sugar. While matched for appearance, maltose could theoretically have its own effects on gut microbiota or mood (e.g., via glycemic response), though this is unlikely at the doses used.

Dropout was 22% (13 of 60), which is high but typical for inpatient psychiatric trials. The ITT analysis attempts to handle this, but missing data can still bias results.

Brain imaging was only done at two time points (baseline and post-intervention), not at follow-up, so it's unknown whether neural changes persist.

**Primary outcome (HAM-D):**

HAM-D scores decreased over time in both groups, but the decrease was significantly larger in the probiotic group than the placebo group.

**Interaction (time × group):** Significant (p = 0.01 in the mITT sample; p = 0.02 in the ITT sample).

**Post-hoc comparison (change from baseline to post-intervention):** Probiotic group showed a mean reduction of ~5.5 points on HAM-D; placebo group showed ~2.5 points reduction. The between-group difference was approximately 3 points on HAM-D (Cohen's d = 0.6, a medium effect size).

**Follow-up (week 8):** The difference between groups was maintained at the 4-week follow-up, with the probiotic group still showing lower HAM-D scores than placebo (p = 0.04).

**Secondary clinical outcomes:**

**BDI (self-reported depression):** No significant difference between groups (p > 0.05). The probiotic group showed a trend toward improvement, but it was not statistically significant.

**STAI-1 (anxiety):** No significant difference between groups.

**GSRS (gastrointestinal symptoms):** No significant difference between groups.

**Gut microbiota:**

**Microbial diversity:** The probiotic group maintained microbial diversity (Shannon index) from baseline to post-intervention, while the placebo group showed a significant decrease in diversity (p = 0.03). This suggests probiotics prevented a loss of diversity that occurred in the placebo group (possibly due to hospitalization or medication).

**Lactobacillus abundance:** The probiotic group showed a significant increase in the relative abundance of *Lactobacillus* (the genus containing several of the probiotic strains) compared to placebo (p = 0.01). This increase was correlated with the decrease in HAM-D scores (r = -0.45, p = 0.04), meaning that people who had the biggest increase in *Lactobacillus* also had the biggest improvement in depression.

**Enterotypes:** No significant shift in enterotypes (Prevotella vs. Bacteroides 2) was observed between groups.

**Other taxa:** No significant changes in other bacterial genera (e.g., Bifidobacterium, Streptococcus) were detected.

**Brain imaging (fMRI):**

**Putamen activation:** The probiotic group showed a significant decrease in putamen activation in response to neutral faces from baseline to post-intervention, compared to the placebo group (p = 0.02, corrected for multiple comparisons). The putamen is a brain region involved in reward processing and motor function.

**Amygdala activation:** No significant difference between groups in amygdala response to fearful, happy, or neutral faces. This contradicts the authors' hypothesis (based on a prior IBS study).

**Other regions:** No significant changes in caudate or nucleus accumbens activation.

**Depressive symptoms:** The probiotic group improved by about 5.5 points on the HAM-D (0–52 scale), while the placebo group improved by about 2.5 points. The additional benefit of probiotics was roughly 3 points — equivalent to moving from "moderate" depression (HAM-D 14–18) to "mild" depression (HAM-D 8–13) for some patients. This is a modest but clinically meaningful effect, comparable to the difference between a low-dose antidepressant and placebo in some trials.

**Gut microbiota:** The increase in *Lactobacillus* was specific and detectable — about a 2-fold increase in relative abundance in the probiotic group. The preservation of microbial diversity (Shannon index) was a prevention of decline rather than an absolute increase.

**Brain activation:** The reduction in putamen activation to neutral faces was about a 15–20% decrease in the blood-oxygen-level-dependent (BOLD) signal in the probiotic group, relative to placebo. This is a moderate effect size (Cohen's d ≈ 0.5–0.6). For context, neutral faces are ambiguous social stimuli; reduced putamen activation might indicate less "salience" or "emotional processing" being assigned to ambiguous social cues.

**Small sample size:** Only 47 completers; the study was powered to detect a medium effect (f = 0.25) but not small effects or subgroup differences.

**Short intervention:** 31 days is brief for a psychiatric intervention; longer treatment might produce larger or more durable effects.

**No stand-alone treatment:** All patients were on treatment-as-usual (antidepressants, psychotherapy). The study cannot determine if probiotics alone would work.

**Industry funding:** The probiotic was provided by the manufacturer (Mendes SA). While the authors state the funder had no role in study design or analysis, industry-funded probiotic trials tend to show larger effects than independent trials.

**Change in inclusion criteria:** The relaxation from HAM-D >24 to HAM-D >7 mid-study is a protocol deviation that weakens the study's internal validity.

**No dietary control after discharge:** Inpatients received a standardized diet, but after discharge (which could happen at any point during the 31 days), diet was uncontrolled. Dietary fiber and fermentable carbohydrates can affect gut microbiota independently of probiotics.

**No assessment of blinding success:** The authors did not report whether patients or clinicians could guess which group they were in. If blinding was broken (e.g., because probiotics caused bloating), this could bias results.

**Multiple comparisons:** Many secondary outcomes (BDI, STAI, GSRS, multiple brain regions, multiple bacterial taxa) were tested without correction for multiple comparisons, increasing the risk of false positives.

**No follow-up brain imaging:** Neural changes were only measured immediately after the intervention, so it's unknown if they persist.

**Population limits:** All patients were inpatients in a single Swiss hospital. Results may not generalize to outpatients, other countries, or people with different diets or gut microbiomes.

For someone running their own n=1 experiment:

**What to test:**

A high-dose multi-strain probiotic supplement containing at least 8 strains (including *Lactobacillus* and *Bifidobacterium* species) at a dose of at least 100 billion CFU/day (the study used 900 billion CFU/day, which is very high — most commercial probiotics are 10–50 billion CFU). A reasonable self-experiment dose would be 50–200 billion CFU/day of a multi-strain product.

Alternatively, test a single-strain *Lactobacillus* or *Bifidobacterium* supplement at 10–50 billion CFU/day to see if you get a similar effect.

**Minimum meaningful duration:**

At least 31 days (4 weeks) based on this study. The effect persisted for 4 weeks after stopping, so you could test a 4-week intervention followed by a 4-week washout.

For a self-experiment, consider a 4-week baseline (no supplement), 4-week intervention, and 4-week follow-up (no supplement) to see if symptoms return.

**What to measure (specific metrics):**

**Primary metric:** Depressive symptoms using a validated self-report scale. The Beck Depression Inventory (BDI) is free and takes 5–10 minutes. Alternatively, the Patient Health Questionnaire (PHQ-9) is widely used and free. Score weekly.

**Secondary metrics:**

- Anxiety (GAD-7, free, weekly)

- Gastrointestinal symptoms (GSRS or a simple stool diary: Bristol Stool Scale, frequency, bloating)

- Sleep quality (Pittsburgh Sleep Quality Index or a simple sleep diary)

- Energy/motivation (visual analog scale 0–10 daily)

**Gut microbiota (optional):** If you have access to a 16S rRNA sequencing service (e.g., uBiome, Thryve, or a research lab), collect stool samples at baseline, end of intervention, and end of follow-up. Look for changes in *Lactobacillus* abundance and overall diversity (Shannon index).

**Brain activity (optional):** Not feasible for most self-experimenters, but you could track reaction time or emotional processing using online cognitive tests (e.g., emotional Stroop test, facial emotion recognition tasks).

**Key confounds to control for:**

**Medication:** If you are on antidepressants, do not change your dose during the experiment. Record any changes.

**Diet:** Keep your diet as consistent as possible, especially fiber intake (preb