Study analysis · Scientific Reports · 2026
Your gut bacteria might predict how fast you're aging—and one species stands out.
In a small study, certain gut bacteria were linked to a measure of aging speed, but it's not proof they cause aging.
Overview
What the study found
The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.
In simple terms
This study looked at whether the bacteria in your gut are linked to how fast your body ages. They found some bacteria that seem to go along with faster or slower aging, but because they only checked one moment in time, we can't say that the bacteria cause aging—just that they are related.
What’s the bottom line?
Scientists looked at gut bacteria and blood aging marks from 123 people. They asked whether gut bacteria can predict how fast someone is biologically aging.
How strong is this study?
The study is a decent first step, but it has some weaknesses. It only included 123 people from one specific group, and it can't prove cause and effect. We need more research with more people over time to be sure about the results.
40 / 100
- COI disclosure+40/40
- Data availabilitydata not shared
- Code availabilitycode not shared
11 / 100
- Randomizationnot randomized
- Blindingblinding unclear
- Control groupno control group
- Sample size (n=123)+9.2/20
- Follow-upno follow-up reported
100 / 100
54 / 100
- P-values+15/15
- Effect size+20/20
- Confidence intervalsno confidence intervals
- Pre-registrationnot pre-registered
Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.
Where it sits
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 544 / 100
Probability of being correct
Snapshots of a population at a single point in time, or descriptions of small groups. Can identify correlations and prevalence, but cannot determine cause and effect.
This design cannot establish causation — the findings describe an association, not a cause. Cross-sectional observational design with no randomization, no control group, and no temporal sequence. Associations may be confounded by diet, BMI, age, medications, and other lifestyle factors. The study cannot determine whether microbiome changes cause epigenetic aging or whether aging-related changes alter the microbiome.
COI Unknown
Could not determine conflict of interest status
No conflict of interest or funding statement was present in the provided text, so severity cannot be determined.
The provided excerpt does not include author affiliations, a conflict of interest section, or a funding statement. The study appears to use an academic cohort (HI-SEED), but without disclosure, industry funding and author COIs cannot be assessed.
Key takeaways
- 01
The germ model explained about 15% of the variance in DunedinPACE aging pace at species level (R²=0.152, relative variance explained; permutation p<0.001) and about 10% at genus level (R²=0.099; permutation p=0.036).
- 02
It did not predict other aging clocks.
- 03
Bifidobacterium adolescentis was linked to slower aging; Succinivibrio dextrinosolvens was linked to faster aging.
- 04
No absolute risk increase was reported because this was not a disease-risk study.
- 05
This does not mean gut bacteria cause aging or that changing them adds years.
- 06
The absolute risk increase was not reported in this study.
- 07
A 15% variance explained is modest, meaning other factors likely matter more.
- 08
The study cannot say how many people would be affected.
Surprising findings
- Traditional epigenetic clocks (Horvath, Levine, GrimAge2) showed no predictive signal from the microbiome, while DunedinPACE did.DunedinPACE is a newer clock that measures pace of aging, not cumulative age. The fact that it uniquely correlates with the microbiome suggests it captures a dynamic aspect of aging that traditional clocks miss.
- Bifidobacterium adolescentis was the dominant predictor of slower aging, despite not being among the top taxa in univariate correlations.This indicates suppressor variable dynamics: its importance emerges only when considering other bacteria, meaning its effect is unique and not captured by simple correlations.
- Succinivibrio dextrinosolvens, a succinate producer, was associated with faster aging, contrary to the typical view that SCFA producers are beneficial.It challenges the simplistic narrative that all short-chain fatty acid producers are good for health and aging.
Practical takeaways
Focus on overall metabolic health (maintain healthy blood sugar and weight) as it correlates with slower aging pace.
This is correlational; improving metabolic health may or may not directly slow aging pace, but it's beneficial for many reasons.
medium confidenceConsider eating a fiber-rich diet that supports beneficial gut bacteria like Bifidobacterium, but don't expect it to dramatically slow aging based on this study alone.
The study doesn't prove that increasing B. adolescentis slows aging. Probiotic supplements may not have the same effect as naturally occurring bacteria.
low confidenceDon't rely on microbiome tests to predict your aging speed; they are not diagnostic tools yet.
The study's models are not intended for individual-level prediction and explain only ~15% of variance.
high confidenceWhy this study matters
Gut Bacteria Predict Aging Speed (But Only One Clock)
In 123 adults, the gut microbiome predicted DunedinPACE, a DNA methylation-based pace-of-aging measure, with a held-out R² of 0.152 at species level (15.2% variance explained, a relative measure) and 0.099 at genus level. However, it showed no predictive signal for traditional epigenetic clocks (Horvath, Levine, GrimAge2; all permutation p>0.11).
This suggests that the gut microbiome may be more closely linked to the rate of biological aging than to cumulative age estimates, offering a new angle on how lifestyle factors like diet could influence aging.
Age-Independent Signal: Not Just Old People Have Different Bacteria
Adding chronological age to the microbiome model did not improve prediction of DunedinPACE (species ΔR² = −0.046; genus ΔR² = −0.005), and key species like Bifidobacterium adolescentis showed no correlation with chronological age. This indicates the association is not merely due to age-related microbial shifts.
Many aging studies struggle to separate biological aging from chronological age. This finding suggests the microbiome captures something about aging pace that is independent of how many years you've lived.
Bifidobacterium adolescentis: The Anti-Aging Bug?
Bifidobacterium adolescentis was the dominant predictor of decelerated aging (mean SHAP = −0.007), with a mean absolute SHAP value 2.5 times larger than the next feature. It has been linked to anti-inflammatory properties and folate/GABA production.
This specific species is already available in some probiotics, so people might wonder if they should take it. But the study is correlational—it doesn't prove that increasing B. adolescentis slows aging.
Succinivibrio dextrinosolvens: The Aging Accelerator?
Succinivibrio dextrinosolvens showed the strongest positive association with accelerated aging (mean SHAP = +0.006). Interestingly, it's a succinate producer, which typically is considered beneficial, creating a paradox.
This challenges the idea that all short-chain fatty acid producers are good. It highlights the complexity of the microbiome and how little we know about specific taxa.
Metabolic Health Tracks with Faster Aging Pace
Higher DunedinPACE was associated with higher HbA1c (r=0.20, p=0.026), higher BMI (r=0.31, p<0.001), and was elevated in type 2 diabetes (1.31 vs 1.18, p=0.050) and obesity (1.27 vs 1.12, p<0.001).
This reinforces the link between metabolic health and aging, but it's cross-sectional—we can't tell if poor metabolic health accelerates aging or vice versa.
No Overall Community Shift, Just Specific Taxa
Despite strong taxon-level associations, overall gut community composition did not differ between extreme high vs low DunedinPACE groups (PERMANOVA p>0.17). This suggests a distributed multivariable signature rather than a broad shift.
It means you can't just look at someone's overall microbiome diversity to predict aging speed; you need to look at specific bacteria and their combined effects.
Want the whole report?
Detailed mode opens the full scientific breakdown — every score component, the methodology, conflicts of interest, the evidence analysis behind each claim, and the raw study data.
Overview
What the study found
The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.
Not medical advice. For informational purposes only. Always consult a healthcare professional. Terms
Scientists looked at gut bacteria and blood aging marks from 123 people. They asked whether gut bacteria can predict how fast someone is biologically aging.
Research results
The germ model explained about 15% of the variance in DunedinPACE aging pace at species level (R²=0.152, relative variance explained; permutation p<0.001) and about 10% at genus level (R²=0.099; permutation p=0.036). It did not predict other aging clocks. Bifidobacterium adolescentis was linked to slower aging; Succinivibrio dextrinosolvens was linked to faster aging. No absolute risk increase was reported because this was not a disease-risk study.
What this means - more context
This does not mean gut bacteria cause aging or that changing them adds years. The absolute risk increase was not reported in this study. A 15% variance explained is modest, meaning other factors likely matter more. The study cannot say how many people would be affected.
Proof-of-concept cross-sectional study testing whether gut microbiome composition predicts DNA methylation-based biological aging metrics independent of chronological age.
In 123 adults aged 17–82, paired stool 16S rRNA and monocyte DNA methylation data were used. Microbiome models predicted DunedinPACE, a pace-of-aging measure, at species level (held-out R²=0.152; permutation p<0.001) and genus level (R²=0.099; permutation p=0.036), but showed no predictive signal for Horvath, Levine, or GrimAge2 acceleration residuals (all permutation p>0.11). Bifidobacterium adolescentis was the dominant predictor of decelerated aging; Succinivibrio dextrinosolvens was the strongest predictor of accelerated aging. Adding chronological age did not improve DunedinPACE prediction (species ΔR²=−0.046; genus ΔR²=−0.005). No absolute risk increase was reported or derivable because outcomes are continuous aging biomarkers, not disease events.
Methods Used
Cross-sectional cohort of 123 adults from the HI-SEED study, enriched for Native Hawaiian and Pacific Islander participants. Same-day stool and blood collection; 16S rRNA gene sequencing and Illumina EPIC methylation arrays on monocyte-enriched samples. Epigenetic clocks included Horvath, Levine, GrimAge2, and DunedinPACE. Machine learning used ElasticNet, SVM, XGBoost, and deep neural networks with 70/30 train-test split, fivefold cross-validated Bayesian hyperparameter optimization, and 1,000 permutation tests. SHAP analysis was restricted to the best species-level DunedinPACE ElasticNet model.
Main Finding
Gut microbiome composition predicted DunedinPACE at species level (held-out R²=0.152; Spearman ρ=0.408, p=0.012; permutation p<0.001) and genus level (R²=0.099; permutation p=0.036). Adding chronological age did not improve prediction (species ΔR²=−0.046; genus ΔR²=−0.005), supporting age-independence. Traditional clock residuals showed no predictive utility (all permutation p>0.11). Bifidobacterium adolescentis was the dominant contributor and strongest predictor of decelerated aging (mean SHAP=−0.007); Succinivibrio dextrinosolvens was the strongest predictor of accelerated aging (mean SHAP=+0.006). These are variance-explained and correlational effects, not absolute risk increases; the absolute risk increase was not reported in this study.
Confidence Level
Low-to-moderate. Strengths include paired same-day sampling, permutation testing, and multiple machine learning algorithms. Limitations include small sample size (N=123; held-out test N=37), cross-sectional design precluding causal inference, no external validation, 16S resolution limits at species level, NHPI-enriched cohort may limit generalizability, and no formal correction across all four clocks. Findings are explicitly hypothesis-generating.
Study Flags
Red Flags
- •Small sample (N=123; held-out test N=37) limits precision and power
- •Cross-sectional design cannot establish causality or directionality
- •No external validation; 16S resolution and NHPI-enriched cohort may limit generalizability; no correction across all four clocks
No biological mechanisms were identified in this study. This may be an epidemiological, observational, or survey-based study that reports associations rather than proposing causal biological pathways.
Surprising Findings
Traditional epigenetic clocks (Horvath, Levine, GrimAge2) showed no predictive signal from the microbiome, while DunedinPACE did.
DunedinPACE is a newer clock that measures pace of aging, not cumulative age. The fact that it uniquely correlates with the microbiome suggests it captures a dynamic aspect of aging that traditional clocks miss.
Practical Takeaways
Focus on overall metabolic health (maintain healthy blood sugar and weight) as it correlates with slower aging pace.
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 544 / 100
Probability of being correct
Snapshots of a population at a single point in time, or descriptions of small groups. Can identify correlations and prevalence, but cannot determine cause and effect.
Human Cross-Sectional
Subject
Moderate probability
on the GRADE evidence scale
This study looked at whether the bacteria in your gut are linked to how fast your body ages. They found some bacteria that seem to go along with faster or slower aging, but because they only checked one moment in time, we can't say that the bacteria cause aging—just that they are related.
Strengths
- Paired stool and blood samples collected at the same time point.
- Use of multiple epigenetic clocks including DunedinPACE.
- Application of machine learning with cross-validation and permutation testing.
Weaknesses
- Cross-sectional design cannot establish causality.
- Modest sample size (N=123) limits statistical power and generalizability.
- No external validation in independent cohorts.
Methodology
Evidence Keywords
Statistical Reporting
Not medical advice. For informational purposes only. Always consult a healthcare professional. Terms
Scientists looked at gut bacteria and blood aging marks from 123 people. They asked whether gut bacteria can predict how fast someone is biologically aging.
Research results
The germ model explained about 15% of the variance in DunedinPACE aging pace at species level (R²=0.152, relative variance explained; permutation p<0.001) and about 10% at genus level (R²=0.099; permutation p=0.036). It did not predict other aging clocks. Bifidobacterium adolescentis was linked to slower aging; Succinivibrio dextrinosolvens was linked to faster aging. No absolute risk increase was reported because this was not a disease-risk study.
What this means - more context
This does not mean gut bacteria cause aging or that changing them adds years. The absolute risk increase was not reported in this study. A 15% variance explained is modest, meaning other factors likely matter more. The study cannot say how many people would be affected.
Proof-of-concept cross-sectional study testing whether gut microbiome composition predicts DNA methylation-based biological aging metrics independent of chronological age.
In 123 adults aged 17–82, paired stool 16S rRNA and monocyte DNA methylation data were used. Microbiome models predicted DunedinPACE, a pace-of-aging measure, at species level (held-out R²=0.152; permutation p<0.001) and genus level (R²=0.099; permutation p=0.036), but showed no predictive signal for Horvath, Levine, or GrimAge2 acceleration residuals (all permutation p>0.11). Bifidobacterium adolescentis was the dominant predictor of decelerated aging; Succinivibrio dextrinosolvens was the strongest predictor of accelerated aging. Adding chronological age did not improve DunedinPACE prediction (species ΔR²=−0.046; genus ΔR²=−0.005). No absolute risk increase was reported or derivable because outcomes are continuous aging biomarkers, not disease events.
Methods Used
Cross-sectional cohort of 123 adults from the HI-SEED study, enriched for Native Hawaiian and Pacific Islander participants. Same-day stool and blood collection; 16S rRNA gene sequencing and Illumina EPIC methylation arrays on monocyte-enriched samples. Epigenetic clocks included Horvath, Levine, GrimAge2, and DunedinPACE. Machine learning used ElasticNet, SVM, XGBoost, and deep neural networks with 70/30 train-test split, fivefold cross-validated Bayesian hyperparameter optimization, and 1,000 permutation tests. SHAP analysis was restricted to the best species-level DunedinPACE ElasticNet model.
Main Finding
Gut microbiome composition predicted DunedinPACE at species level (held-out R²=0.152; Spearman ρ=0.408, p=0.012; permutation p<0.001) and genus level (R²=0.099; permutation p=0.036). Adding chronological age did not improve prediction (species ΔR²=−0.046; genus ΔR²=−0.005), supporting age-independence. Traditional clock residuals showed no predictive utility (all permutation p>0.11). Bifidobacterium adolescentis was the dominant contributor and strongest predictor of decelerated aging (mean SHAP=−0.007); Succinivibrio dextrinosolvens was the strongest predictor of accelerated aging (mean SHAP=+0.006). These are variance-explained and correlational effects, not absolute risk increases; the absolute risk increase was not reported in this study.
Confidence Level
Low-to-moderate. Strengths include paired same-day sampling, permutation testing, and multiple machine learning algorithms. Limitations include small sample size (N=123; held-out test N=37), cross-sectional design precluding causal inference, no external validation, 16S resolution limits at species level, NHPI-enriched cohort may limit generalizability, and no formal correction across all four clocks. Findings are explicitly hypothesis-generating.
Study Flags
Red Flags
- •Small sample (N=123; held-out test N=37) limits precision and power
- •Cross-sectional design cannot establish causality or directionality
- •No external validation; 16S resolution and NHPI-enriched cohort may limit generalizability; no correction across all four clocks
No biological mechanisms were identified in this study. This may be an epidemiological, observational, or survey-based study that reports associations rather than proposing causal biological pathways.
Surprising Findings
Traditional epigenetic clocks (Horvath, Levine, GrimAge2) showed no predictive signal from the microbiome, while DunedinPACE did.
DunedinPACE is a newer clock that measures pace of aging, not cumulative age. The fact that it uniquely correlates with the microbiome suggests it captures a dynamic aspect of aging that traditional clocks miss.
Practical Takeaways
Focus on overall metabolic health (maintain healthy blood sugar and weight) as it correlates with slower aging pace.
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 544 / 100
Probability of being correct
Snapshots of a population at a single point in time, or descriptions of small groups. Can identify correlations and prevalence, but cannot determine cause and effect.
Human Cross-Sectional
Subject
Moderate probability
on the GRADE evidence scale
This study looked at whether the bacteria in your gut are linked to how fast your body ages. They found some bacteria that seem to go along with faster or slower aging, but because they only checked one moment in time, we can't say that the bacteria cause aging—just that they are related.
Strengths
- Paired stool and blood samples collected at the same time point.
- Use of multiple epigenetic clocks including DunedinPACE.
- Application of machine learning with cross-validation and permutation testing.
Weaknesses
- Cross-sectional design cannot establish causality.
- Modest sample size (N=123) limits statistical power and generalizability.
- No external validation in independent cohorts.
Methodology
Evidence Keywords
Statistical Reporting
Scoring
How strong is this study?
The study is a decent first step, but it has some weaknesses. It only included 123 people from one specific group, and it can't prove cause and effect. We need more research with more people over time to be sure about the results.
40 / 100
- COI disclosure+40/40
- Data availabilitydata not shared
- Code availabilitycode not shared
11 / 100
- Randomizationnot randomized
- Blindingblinding unclear
- Control groupno control group
- Sample size (n=123)+9.2/20
- Follow-upno follow-up reported
100 / 100
54 / 100
- P-values+15/15
- Effect size+20/20
- Confidence intervalsno confidence intervals
- Pre-registrationnot pre-registered
Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.
Where it sits
RCT reviewsReviews of RCTs (Meta-analyses)
Max 100Randomized TrialsRandomized Trials
Max 90Reviews of Cohort StudiesReviews of Cohort Studies
Max 85Cohort StudiesCohort Studies
Max 72Reviews of Case-Control StudiesReviews of Case-Control Studies
Max 63Case-Control StudiesCase-Control Studies
Max 58Cross-Sectional & Case SeriesCross-Sectional & Case Series
Max 50Expert OpinionExpert Opinion
Max 544 / 100
Probability of being correct
Snapshots of a population at a single point in time, or descriptions of small groups. Can identify correlations and prevalence, but cannot determine cause and effect.
This design cannot establish causation — the findings describe an association, not a cause. Cross-sectional observational design with no randomization, no control group, and no temporal sequence. Associations may be confounded by diet, BMI, age, medications, and other lifestyle factors. The study cannot determine whether microbiome changes cause epigenetic aging or whether aging-related changes alter the microbiome.
COI Unknown
Could not determine conflict of interest status
No conflict of interest or funding statement was present in the provided text, so severity cannot be determined.
The provided excerpt does not include author affiliations, a conflict of interest section, or a funding statement. The study appears to use an academic cohort (HI-SEED), but without disclosure, industry funding and author COIs cannot be assessed.
Standing
The people behind it
The researchers who wrote the study this analysis is built on.
Authored by
4 researchersIf this is your work, this is how we attribute it on Fit Body Science. Braden P. Kunihiro is listed as the lead author.
- Alika K. MaunakeaCorresponding