Meta-analysis of probiotic/synbiotic interventions in metabolic syndrome: preclinical and clinical evidence
Probiotics/synbiotics improve body weight, waist circumference, and lipid profile in metabolic syndrome patients, but do not significantly reduce fasting plasma glucose in humans.
| Population | Patients with metabolic syndrome (12 clinical trials, n=696) and animal models of MetS (23 studies) |
|---|---|
| Intervention | Probiotic or synbiotic supplementation |
| Comparator | Placebo or no-intervention control |
| Outcome | Triglycerides (humans); Total cholesterol (humans); LDL-C (humans); HDL-C (humans); Body weight (humans); Waist circumference (humans); Fasting plasma glucose (humans) |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| Triglycerides (humans) | MD -20.51 mg/dL, 95% CI -27.69 to -13.34 | — | Moderate | — | 12 studies |
| Total cholesterol (humans) | MD -12.09 mg/dL, 95% CI -16.15 to -8.04 | — | Moderate | — | 12 studies |
| LDL-C (humans) | MD -10.56 mg/dL, 95% CI -14.96 to -6.17 | — | Moderate | — | 12 studies |
| HDL-C (humans) | MD +1.97 mg/dL, 95% CI 0.55 to 3.39 | — | Moderate | — | 12 studies |
| Body weight (humans) | MD -1.06 kg, 95% CI -1.70 to -0.42 | — | Moderate | — | 12 studies |
| Waist circumference (humans) | MD -2.12 cm, 95% CI -3.70 to -0.55 | — | Moderate | — | 12 studies |
| Fasting plasma glucose (humans) | MD -8.76 mg/dL, 95% CI -18.29 to 0.77 | — | Moderate | — | 12 studies |
Context
Metabolic syndrome affects hundreds of millions worldwide and substantially raises cardiovascular and type 2 diabetes risk. Prior meta-analyses on probiotics in MetS showed inconsistent results and limited mechanistic exploration. This study integrated preclinical and clinical data to quantify metabolic and inflammatory effects.
What the study showed
In patients, intervention reduced body weight by 1.06 kg (95% CI: –1.70 to –0.42), waist circumference by 2.12 cm (95% CI: –3.70 to –0.55), triglycerides by 20.51 mg/dL (95% CI: –27.69 to –13.34), total cholesterol by 12.09 mg/dL (95% CI: –16.15 to –8.04), LDL-C by 10.56 mg/dL (95% CI: –14.96 to –6.17), and raised HDL-C by 1.97 mg/dL (95% CI: 0.55 to 3.39). Fasting plasma glucose change in humans was non-significant (MD = –8.76 mg/dL, 95% CI: –18.29 to 0.77). In animal models, all outcomes including glucose, HbA1c, and inflammatory markers were significantly improved.
How it was done
Meta-analysis of 35 studies (12 clinical trials with 696 patients and 23 animal studies), searching PubMed, Embase, Cochrane CENTRAL, CNKI, and Wanfang Data through June 2025. Analyses performed in Stata 15.0 computing pooled mean differences (MD) with 95% CIs; sensitivity and subgroup analyses assessed stability and heterogeneity.
Effect magnitude
The strongest clinical effect was for triglycerides (MD = –20.51 mg/dL, 95% CI: –27.69 to –13.34) and total cholesterol (MD = –12.09 mg/dL, 95% CI: –16.15 to –8.04); these magnitudes are modest relative to typical baseline MetS values.
Risk of bias
The clinical sample is small (n=696 across 12 trials), limiting precision and generalizability. Simultaneous pooling of animal and human data introduces substantial methodological and biological heterogeneity. Risk of bias assessment tools (RoB 2, AMSTAR-2) are not described with sufficient detail in the available text; heterogeneity across strains, doses, and durations was not fully resolved by subgroup analyses.
What this study does NOT prove
This study does not prove that probiotics prevent cardiovascular events or incident diabetes in MetS patients. Animal data are not directly translatable to humans, and the small clinical sample precludes definitive conclusions on long-term safety and efficacy.
In clinical practice
Probiotics/synbiotics may be considered adjuncts for modest lipid improvement and weight reduction in MetS patients, without replacing established pharmacological or lifestyle interventions. Clinicians should not expect significant fasting glucose reduction based on this evidence. Lack of strain and dose standardization prevents specific product recommendations.
Limitations
The clinical sample is small (n=696 across 12 trials), limiting precision and generalizability. Simultaneous pooling of animal and human data introduces substantial methodological and biological heterogeneity. Risk of bias assessment tools (RoB 2, AMSTAR-2) are not described with sufficient detail in the available text; heterogeneity across strains, doses, and durations was not fully resolved by subgroup analyses.
What is still missing
Long-term RCTs with larger samples, standardized strains and doses, and hard cardiovascular endpoints (events, mortality) are needed to validate clinical relevance.
Technical appendix
Version history
- 1.0 · 2026-10-31 — Auto-generated under Evidence Standard v1.0