← Back

Mbiotix-PD

Safety and Efficacy of Fecal Microbiota Transplantation in Alleviating Symptoms of Parkinson's Disease: A Randomized, Placebo-Controlled, Double-Blinded Study

Share Share Copied! Compare

Year of Publication: 2026

Authors: Figura M, Milanowski Ł, Nowak JM, et al.

Journal: Annals of Neurology

Citation: Ann Neurol 2026;100(1):10–21. DOI: 10.1002/ana.78153

Link: https://doi.org/10.1002/ana.78153

Clinical Question

Does a single donor-derived fecal microbiota transplant, delivered by colonoscopy, improve motor (MDS-UPDRS III “OFF”) and non-motor symptoms of Parkinson disease over 12 months compared with autologous-stool placebo FMT?

Bottom Line

A single colonoscopy-delivered donor FMT did NOT improve the primary motor endpoint (MDS-UPDRS III “OFF”) at 12 months in PD, but produced significant, mostly modest improvements in several non-motor domains (total MDS-NMS, pain, cognition subscale, sexual dysfunction, GI dysfunction) and in EQ-5D quality of life at 12 months. The procedure was safe with only transient GI adverse events.

        Major Points
        Design: prospective, parallel-group, randomized 1:1, double-blind, placebo (auto-FMT)-controlled, single-center trial at Medical University of Warsaw / Mazovian Brodnowski Hospital, Poland; enrollment Jan 2022–Aug 2023; 12-month follow-up.
Population: 59 patients with MDS-criteria PD on oral levodopa, aged 40–75 y (median 65 y Mbiotix / 63 y placebo); 30 female (50.8%), 29 male (49.2%); median symptom duration 5–6 y; baseline LEDD ~796 (Mbiotix) vs ~742 (placebo) mg/day; BMI higher in Mbiotix (27.2 vs 25.8 kg/m², p=0.015).
Intervention: single 60 g donor stool suspended in 200 mL saline with 10% glycerol, injected into cecum via colonoscopy after 5-day rifaximin + bowel prep; comparator: identically prepared autologous (patient's own) stool.
Primary outcome (MDS-UPDRS III “OFF” at 12 mo): DiD +1.50 (95% CI −4.28 to 7.28, p=1.00 Bonferroni-adjusted) — no difference; both arms slightly worsened (∆ +3.0 Mbiotix vs +1.5 placebo).
MDS-UPDRS III “ON” at 1 mo showed early significant benefit (DiD −3.5, 95% CI −5.2 to −1.8, p<0.001) that was not sustained.
Non-motor: MDS-NMS total at 12 mo DiD −20.3 (95% CI −35.2 to −5.4, adj p=0.030); pain subscale reductions at 1 mo (−3.9), 3 mo (−4.9), and 12 mo (−3.6), all p<0.001; cognition subscale at 3 mo DiD −5.5 (p<0.001); sexual dysfunction at 6 mo DiD −1.8 (p<0.001).
GI: GIDS-PD at 12 mo DiD −4.4 (95% CI −6.9 to −1.8, p=0.003) favoring FMT; mCAS unchanged.
Quality of life: EQ-5D at 12 mo DiD +0.03 (95% CI 0.01–0.05, p=0.037) favoring FMT (placebo group declined); PDQ-39 no significant difference.
MoCA: within-group improvement in Mbiotix (∆ +0.90, p=0.049) but between-group difference NS (DiD 0.6, 95% CI −0.01 to 1.3, p=0.10).
LEDD unchanged in both arms (12-mo DiD ~−4, p=1.000).
Safety: only transient GI adverse events (abdominal pain, diarrhea, constipation, heartburn, nausea, painful defecation, fever); no serious procedure-related events; 36/59 (61%) chose general anesthesia for colonoscopy.

      

Design

Study Type: Randomized, parallel-group, double-blind, placebo (autologous-FMT)-controlled trial

Randomization: 1

Blinding: Double-blind (patients and clinicians blinded to allocation; only the head of the HBI Manufacturing Laboratory was unblinded to prepare/label samples). Sealed-Envelope software; 1:1; stratified by PD stage (early vs advanced).

Enrollment Period: January 2022 – August 2023

Follow-up Duration: 12 months (assessments at baseline, 1, 3, 6, and 12 mo; MoCA at baseline and 12 mo)

Centers: 1

Countries: Poland

Sample Size: 59

Power Calculation: Powered on primary DiD estimator at 12 mo (2-sample t-test approximation, Cohen's d = 0.8, 80% power, α=0.05): 21/arm minimum, inflated to 26/arm (52 total) to account for 15–20% attrition.

Analysis: Intention-to-treat using a robust linear mixed-effects model with treatment, time, treatment×time, age, sex, BMI as fixed effects and group-level random effects; robust estimation; predictive-mean-matching imputation for missing data (assumed MNAR); DiD estimator; Bonferroni adjustment across 4 timepoints (adjusted α=0.0125); effect sizes as Cohen's d.

Inclusion Criteria

Diagnosis of Parkinson disease per MDS clinical diagnostic criteria
Ongoing oral levodopa treatment
Age 40–75 years
Written informed consent

Exclusion Criteria

Contraindications for colonoscopy (history of gastrointestinal perforation or obstruction)
Post-radiotherapy status of the abdominal or rectal area
Severe heart, kidney, or liver failure
Coagulation disorders
Congenital or acquired immune deficiencies
Current viral, bacterial, or fungal infection
Abdominal aortic aneurysm requiring surgery
Pregnancy or lactation
Treatment with dopaminergic infusion therapies
Deep brain stimulation
Colorectal polyps at colonoscopy (except lesions <5 mm classified as NBI International Colorectal Endoscopic type I) or other potentially neoplastic changes
Strong food allergy with history of anaphylaxis
Stool positive for MDR/pathogenic bacteria (MRSA, KPC, MBL- or ESBL-producing organisms, VRE, Salmonella, Shigella, Yersinia)

Arms

Field	Mbiotix (donor FMT)	Control
Intervention	Single donor-derived FMT: 60 g SuperDonor stool in 200 mL 0.9% NaCl with 10% glycerol (Mbiotix, Human Biome Institute), stored at −80°C and thawed at room temperature, delivered into cecum via colonoscopy after 5-day rifaximin 400 mg PO tid + bowel prep with macrogol 4000.	Single autologous FMT: identically prepared 60 g suspension of the patient's own stool (collected before antibiotic pretreatment) in 200 mL 0.9% NaCl with 10% glycerol, delivered into cecum via colonoscopy after 5-day rifaximin 400 mg PO tid + bowel prep.
N	28	31

Outcomes

Outcome	Type	Control	Intervention	P-value
Change in MDS-UPDRS Part III (motor examination) “OFF” state score from baseline to 12 months, Mbiotix vs placebo (auto-FMT).	Primary	12-mo EMM 24.0 (95% CI 19.9–28.1); change from baseline +1.5 (95% CI −0.6 to 3.7)	12-mo EMM 24.5 (95% CI 20.6–28.4); change from baseline +3.0 (95% CI 1.0 to 5.0)	1.00 (Bonferroni-adjusted across 4 timepoints)
MDS-UPDRS III “ON” at 1 mo (DiD)	Secondary	∆ +1.5 [+0.3, +2.7]	∆ −2.0 [−3.2, −0.8]	<0.001
MDS-UPDRS III “ON” at 12 mo (DiD)	Secondary	∆ +1.6	∆ +0.4	0.168
MoCA at 12 mo (DiD)	Secondary	∆ +0.30	∆ +0.90	0.10 (adjusted)
MDS-UPDRS I at 12 mo (DiD)	Secondary	∆ +0.6	∆ −0.1	0.028
MDS-UPDRS II at 12 mo (DiD)	Secondary	∆ 0.0	∆ −1.0	<0.001
MDS-UPDRS IV at 12 mo (DiD)	Secondary	∆ −0.5	∆ −0.3	1.000
MDS-NMS total at 12 mo (DiD)	Secondary	∆ −1.2	∆ −21.5	0.030
MDS-NMS cognition subscale at 3 mo (DiD)	Secondary	∆ −0.1	∆ −5.5	<0.001
MDS-NMS sexual dysfunction at 6 mo (DiD)	Secondary	∆ +0.7	∆ −1.1	<0.001
MDS-NMS pain at 1 mo (DiD)	Secondary	∆ −0.2	∆ −4.1	<0.001
MDS-NMS pain at 3 mo (DiD)	Secondary	∆ +1.8	∆ −3.2	<0.001
MDS-NMS pain at 12 mo (DiD)	Secondary	∆ 0.0	∆ −3.6	<0.001
MDS-NMS 'other' (fatigue/weight/sweating/olfaction) at 12 mo (DiD)	Secondary	∆ +2.76	∆ −2.9	<0.001
GIDS-PD total at 12 mo (DiD)	Secondary	∆ +1.2	∆ −3.2	0.003
mCAS at 12 mo (DiD)	Secondary	∆ −0.2	∆ −0.8	0.768
PDQ-39 at 12 mo (DiD)	Secondary	∆ +3.6	∆ −0.8	0.296
EQ-5D at 12 mo (DiD)	Secondary	∆ −0.03	∆ 0.00	0.037
LEDD (mg) at 12 mo (DiD)	Secondary	∆ +18	∆ +67	1.000
Any procedure-related adverse event	Safety	same profile	transient GI symptoms only (abdominal pain, diarrhea, constipation, heartburn, epigastric fullness, nausea, painful defecation, fever)
Serious adverse events attributed to procedure	Safety	0/31	0/28
Colonoscopy performed under general anesthesia	Safety	16/31 (52%)	19/28 (68%)
Discontinuations	Safety	—	1 randomized patient excluded intra-procedure for pneumatosis cystoides intestinalis (before FMT); 1 patient lost to follow-up after visit 1

Subgroup Analysis

Randomization was stratified by PD stage (early vs advanced), but no formal subgroup analyses of primary/secondary outcomes by stage, sex, or symptom duration are reported. Authors discuss that Bruggemann et al. (GUT-PARFECT) recruited only early-stage patients (shorter disease duration, lower LEDD) and reported motor benefit, suggesting a possible early-PD-specific effect that this mixed cohort could not detect.

Criticisms

Small sample (N=59) with wide confidence intervals — the primary MDS-UPDRS III OFF DiD 95% CI (−4.28 to 7.28) spans clinically important benefit and harm, so a modest true effect cannot be excluded (type II error risk acknowledged by authors).
Single-center trial in Warsaw, Poland — limited generalizability.
Mixed early and advanced PD population; higher baseline LEDD (~750–800 mg) than GUT-PARFECT's early-PD cohort, which is where prior FMT motor benefit was seen — the negative primary result may reflect stage selection.
Single FMT dose — authors themselves hypothesize that repeated dosing may be required to sustain the early MDS-UPDRS III 'ON' benefit and to produce OFF-state effects.
Multiple secondary endpoints with Bonferroni adjustment only across 4 timepoints (not across scales) — the many positive non-motor subscale findings should be considered exploratory / hypothesis-generating.
No detailed microbiome baseline typing or body-first vs brain-first phenotyping; abnormal baseline microbiota was not an inclusion criterion, unlike Scheperjans's cohort.
Baseline BMI imbalance (higher in Mbiotix, p=0.015) despite randomization — handled via covariate adjustment in the mixed model but residual confounding possible.
Blinding integrity of autologous vs donor FMT not formally assessed (both products identically packaged and stored, but no post-trial guess questionnaire reported).
Sponsor conflict of interest: last-author J. Biliński is a shareholder of Human Biome Institute, the manufacturer of Mbiotix.

Funding

Biocodex Microbiota Foundation National Grant 2021; Medical University of Warsaw grants 09/M/MBS/N/21, 1/Z/MG/24, NZP/1/Z/MG/N/23, and M/MG/77/22. Funders had no role in design or interpretation. Mbiotix product supplied by Human Biome Institute (Warsaw, Poland).

Based on: Mbiotix-PD (Annals of Neurology, 2026)