AIM-MET: AI-Guided Microbiome-Targeted Nutrition for Glycemic Improvement in Type 2 Diabetes

Type 2 diabetes is commonly managed with lifestyle measures and glucose-lowering medications, but many individuals continue to have residual hyperglycaemia despite stable treatment. Nutritional interventions that target the gut microbiome may provide a safe adjunctive approach if they can improve glycaemic control without replacing standard diabetes care.

The gut microbiome may influence glucose regulation through several biological pathways, including short-chain fatty acid production, bile acid metabolism, intestinal barrier function, low-grade inflammation, incretin signalling, and other host-microbe interactions. These mechanisms provide the rationale for evaluating a microbiome-targeted nutritional intervention in adults with type 2 diabetes.

The investigational product is a fixed AI-guided microbiome-targeted nutritional formulation. The artificial intelligence methodology was used before the clinical trial to support formulation design, including ingredient and dose selection from candidate microbiome-active compounds. The output of this process is one fixed formulation that is locked before participant enrolment and is identical for all participants assigned to the active intervention arm. The artificial intelligence system is not used during the trial for participant-level prediction, diagnosis, clinical decision-making, dosing, personalization, or formulation adjustment. Therefore, the clinical study evaluates the fixed nutritional product, not the independent clinical performance of an artificial intelligence system.

Participants are randomized in a 1:1 ratio to receive either the active nutritional intervention or a matching placebo for 24 weeks. Both groups receive identical structured lifestyle counselling and continue stable background glucose-lowering therapy as clinically appropriate. The placebo is matched to the active product in appearance, taste, smell, packaging, administration schedule, and storage conditions, and is intended to be microbiome- and glycaemia-neutral.

The primary clinical assessment is the change in HbA1c from baseline to Week 24. Week 24 was selected because HbA1c reflects average glycaemia over the preceding 2 to 3 months and is therefore appropriate for evaluating a gradually acting nutritional and microbiome-targeted intervention. The study also evaluates supportive glycaemic, anthropometric, safety, patient-reported, and microbiome-related measures.

This 100-participant proof-of-concept trial is designed to detect a clinically meaningful HbA1c signal and to estimate safety, feasibility, variability, and effect-size parameters for a future larger confirmatory trial. The study is not intended to replace standard diabetes care or to assess the investigational product as a substitute for clinically indicated glucose-lowering medication.

Inclusion criteria

Adults aged 18 to 65 years.

Established diagnosis of type 2 diabetes mellitus documented in the medical record, supported by American Diabetes Association diagnostic criteria on at least one prior occasion.

HbA1c 6.8% to 8.2% at screening, measured by an NGSP-certified central laboratory assay; confirmed on a repeat sample if discordant with prior records or if the screening value is at the upper or lower boundary.

Body mass index 18.5 to 40.0 kg/m2 at screening.

Stable background glucose-lowering therapy for at least 3 months before randomization, restricted to metformin, a DPP-4 inhibitor, and/or an SGLT2 inhibitor, either alone or in combination.

Documented body-weight stability, defined as no more than +/-5% or +/-3 kg, whichever is smaller, self-reported body-weight change during the 3 months before screening.

Willing and able to provide written informed consent.

Willing and able to comply with trial visits, study product use, fasting blood sampling, stool sampling, and the study assessment schedule.

Willing to receive standardized lifestyle counselling during the 24-week blinded period and to maintain stable background diabetes management as clinically directed.

No systemic antibiotic use within 8 weeks before randomization.

No probiotic, prebiotic, synbiotic, or postbiotic supplement use within 8 weeks before randomization.

Stable hypertension, dyslipidaemia, and stable thyroid replacement therapy are permitted if medication has been initiated and the dose unchanged for at least 3 months before randomization.

Exclusion criteria

Type 1 diabetes, latent autoimmune diabetes of adults, pancreatogenic diabetes, maturity-onset diabetes of the young, gestational diabetes as the current diagnosis, or any non-type-2 form of diabetes.

HbA1c less than 6.8% or greater than 8.2% at screening.

Current use of GLP-1 receptor agonists or GLP-1/GIP co-agonists, or use within 12 months before randomization.

Current use of basal, prandial, or premixed insulin, or any insulin use within 6 months before randomization.

Current use of sulfonylureas or meglitinides, or use within 3 months before randomization.

Current use of anti-obesity pharmacotherapy or use within 6 months before randomization, including orlistat, naltrexone/bupropion, phentermine/topiramate, or other agents with a primary anti-obesity indication.

Recurrent severe hypoglycaemia or hypoglycaemia unawareness.

Planned initiation or dose escalation of glucose-lowering medication, anti-obesity medication, or systemic corticosteroids during the 24-week trial period in the judgment of the treating clinician.

History of bariatric surgery at any time.

Gastrointestinal surgery other than appendectomy or uncomplicated cholecystectomy.

Active inflammatory bowel disease, coeliac disease, microscopic colitis, chronic pancreatitis, malabsorption syndrome, or chronic severe gastrointestinal disease likely to affect absorption or trial adherence.

Acute gastroenteritis within 4 weeks before randomization.

Colonoscopy bowel preparation within 12 weeks before randomization.

Faecal microbiota transplantation within 12 months before randomization.

Active or recent malignancy within 6 months, except adequately treated non-melanoma skin cancer.

Stage 3b to 5 chronic kidney disease, defined as estimated glomerular filtration rate less than 45 mL/min/1.73 m2.

Decompensated liver disease or Child-Pugh class B/C.

ALT or AST greater than 3 times the upper limit of normal at screening, unless judged clinically insignificant and approved by the investigator.

Known or suspected haemoglobinopathy that materially distorts HbA1c measurement or haemoglobin variant known to interfere with the central laboratory assay.

Blood transfusion, significant blood loss greater than 250 mL, or erythropoiesis-stimulating agent therapy within 3 months before randomization.

Untreated overt thyroid disease, or thyroid medication initiation or dose change within 3 months before randomization.

Use of systemic corticosteroids within 4 weeks before randomization.

Use of immunosuppressive medication or biologic/immunomodulatory therapy within 5 half-lives before randomization.

Pregnancy or lactation; women of childbearing potential not willing to use effective contraception during the 24-week blinded period.

Severe psychiatric illness or cognitive impairment that may compromise informed consent, safety, or adherence.

Active eating disorder, defined operationally as a positive SCOFF screen with at least 2 affirmative responses at screening, or clinical diagnosis of anorexia nervosa, bulimia nervosa, or binge-eating disorder.

Heavy alcohol use or active substance use disorder.

Current very-low-calorie diet, ketogenic diet, medically supervised weight-loss diet, or other highly restrictive diet that excludes major food groups.

Current cigarette smoking with intention to attempt cessation during the 24-week trial period; current users of nicotine-replacement or smoking-cessation pharmacotherapy initiated within 3 months before randomization.

Participation in another interventional clinical trial within 30 days before screening.

Known hypersensitivity or clinically significant intolerance to any component of the active product or placebo.

Inability or unwillingness to provide stool samples.

Any condition that, in the investigator's opinion, would compromise participant safety or trial integrity.